Czech Republic joins ESO

Progress on ALMA
CRIRES Science Verification
Gender balance among ESO staff
The Messenger
No. 128 – June 2007
The Organisation

The Czech Republic Joins ESO

Catherine Cesarsky The XXVIth IAU General Assembly, held This Agreement was formally confirmed
(ESO Director General) in Prague in 2006, clearly provided a by the ESO Council at its meeting on
boost for the Czech efforts to join ESO, 6 December and a few days thereafter by
not the least in securing the necessary the Czech government, enabling a sign-
I am delighted to welcome the Czech public and political support. Thus our ing ceremony in Prague on 22 December
Republic as our 13th member state. From Czech colleagues used the opportunity (see photograph below). This was impor-
its size, the Czech Republic may not be- to publish a fine and very interesting pop- tant because the Agreement foresaw
long to the ‘big’ member states, but the ular book about Czech astronomy and accession by 1 January 2007. With the
accession nonetheless marks an impor- ESO, which, together with the General signatures in place, the agreement could
tant point in ESO’s history and, I believe, Assembly, created considerable media be submitted to the Czech Parliament
in the history of Czech astronomy as well. interest. for ratification within an agreed ‘grace pe-
The Czech Republic is the first of the riod’. This formal procedure was con-
Central and East European countries to On 20 September 2006, at a meeting at cluded on 30 April 2007, when I was noti-
join ESO. The membership underlines ESO Headquarters, the negotiating teams fied of the deposition of the instrument of
ESO’s continuing evolution as the prime from ESO and the Czech Republic arrived ratification at the French Ministry of
European organisation for astronomy, at an agreement in principle, which was Foreign Affairs.
whilst at the same time it enables Czech subsequently presented to the Commit-
astronomers to become fully integrated tee of Council. We look forward to an active Czech parti-
in the European astronomical community. cipation in ESO’s programmes in time to
come.
As always in such cases, joining an inter-
national organisation requires a proc-
ess with a political underpinning and a
set of formal steps. Not surprisingly,
this process can take years. In the case
of the Czech Republic the first informal Ms. Miroslava Kopicová, Minister of
Photo: Academy of Sciences of the Czech Republic

Education, Youth and Sports of

contacts occurred at the time of the
the Czech Republic, and Tom Wilson,
XXVth General Assembly of the IAU in Associate Director of ESO, shake
Sydney, Australia. hands following the signing of the
agreement for the entry of the Czech
Republic into ESO which took place
Discussions continued in the Czech Re-
on 22 December 2006 in Prague.
public leading to an invitation for me to The formal accession took place, with-
visit Prague on 10 November 2005. out ceremony, on 30 April 2007.
During the visit, in which I was accompa-
nied by Claus Madsen, we had very
successful interactions with high-ranking
Czech officials, including Prof. Václav
Pačes, President of the Czech Academy
of Sciences, and representatives of
several ministries. We were also received
by Dr. Martin Jahn, the then Deputy
Prime Minister of the Czech Republic, who
expressed strong support for the Czech
bid to join ESO. The visit also included
a major press conference and a live TV
interview with Dr. Jan Palouš, of the Dr. Jan Palouš of the
Photo: L. Calçada, ESO

Astronomical Institute of the Academy of Astronomical Institute of

the Czech Academy
Sciences, and myself. of Sciences addressing
delegates at the ESO
It was clear that the process was gaining Industry Day held in
considerable momentum, and thus the Prague on 8 June 2007.

formal request for membership arrived in

the spring of 2006, leading to the estab-
lishment by the ESO Council of a negotia-
ting team, chaired by Council Vice-Presi-
dent Dr. Monnik Desmeth.

 The Messenger 128 – June 2007

The Organisation
Astronomy in the Czech Republic
Jan Palouš, Petr Hadrava
(Astronomical Institute, Academy of
Sciences of the Czech Republic)
A brief historical outline of the develop-
ment of astronomy in the Czech lands is
presented, followed by an overview of
the current observational, research and
educational facilities.
Historical development
In the Czech lands, astronomy has rather
deep roots, covering at least all the
last millenium. A significant milestone for
pursuing astronomy in Bohemia and
surrounding countries was represented
by the establishment of the University in
Prague by the Emperor Charles IV in 1348.
Astronomy, as one of the seven liberal
arts, was greatly emphasised right from
the beginning of the lectures, and Prague
University soon began to support the
creative development of astronomical
knowledge brought especially from Paris,
Oxford and other places in Europe. It
reached a remarkable level, e.g. in wide-
ly spread treatises “On the construc-
tion and on theuse of the astrolabe” by
Cristannus de Prachaticz (1407) or in the
design of the Prague Astronomical Clock
(1410) by Johannes Schindel.
The most famous era of astronomy in
Prague was probably the period of rule of
the Emperor Rudolph II at the turn of
the 16th and 17th century, when Tycho
Brahe and Johannes Kepler dwelt and
worked at the Emperor’s court. The fact
that Brahe, and later also Kepler, came
to Prague, as well as Kepler’s success
achieved in this city, were not a result of
some lucky coincidence or the Emperor’s
generosity and eccentric nature. To a
great degree, these were logical conse-
quences of the gradual development
of Prague’s cultural and scientific environ-
ment. It was mainly Tycho’s friend
Thaddaeus Hagecius, who prepared the
grounds for Tycho’s arrival. Hagecius
was the Emperor’s physician, and at the
same time one of a few European as-
tronomers of that time, who, like Tycho
Brahe, studied the skies with a critical
scientific approach. It was in Prague
where Kepler found, during his stay in the
years 1600–1612, his first two laws of
Photo: P. Hadrava, Astronomical Institute, Academy of Sciences of the Czech Republic
Figure 1: The Astrono­mical Clock in the Figure 2: New star of 1572 in the constellation of
Old Town of Prague. Cassiopeia from “Dialexis” by Thaddaeus Hagecius
(Frankfurt am Main 1574).
planetary motions, while evaluating the physiology of the human eye. Of abso-
positions of Mars observed by Tycho lutely cructial importance for the future
Brahe. He published numerous treatises, development of astronomy and other
among them “Astronomia nova” includ- sciences was the idea of the influence of
ing the laws of planetary motions, texts a binary star’s motion on the colour of
on the “New Star” of 1604 in the constel- their light, which Christian Doppler, the
lation Ophiuchus, and he wrote his first professor of the Prague Polytechnic, pre­­-
version of “The Dream, or Posthumous sented in his lecture held in the Caroli-
Work on Lunar astronomy” on the astron- num in 1842. He published his result in a
omy viewed from the surface of the
Moon.
Photo: I. Král, National Library, Prague
After the Thirty Years War, education was
predominantly provided and controlled
by the Jesuit order. Astronomy, as well as
other sciences, was pursued mainly in
the Clementinum College in Prague.
In 1722 the Astronomical Tower was built
there. Thanks to the care of the first direc-
tor of the Clementinum observatory,
Josef Stepling, the Astronomical Tower
was equipped with new instruments
around 1750, becoming the first state-
supported astronomical observatory in
Czech lands.
The advance of astronomy went hand in
hand with the development of other Figure 3: The allegory of sciences on the ceiling
sciences. Among the important scientific fresco in the New Mathematical Hall in Prague
personalities in the 19th century were Clementinum. It shows the Clementinum Astronomi-
cal Tower along with the extrasolar planetary sys-
the mathematician Bernard Bolzano and
tems according to Giordano Bruno’s idea of the
the biologist Jan Evangelista Purkyně, “multiplicity of worlds” (National Library of the Czech
who studied, besides other things, the Republic).
The Messenger 128 – June 2007 
The Organisation Palouš J. and Hadrava P., Astronomy in the Czech Republic

local Bohemian journal, consequently his
principle was later rediscovered by H. A.
Fizeau. In the second half of the 19th cen-
tury, Ernst Mach contributed enormous-
ly to the education of a whole generation
of both German and Czech physicists
and astronomers. He also helped in the
general recognition of Christian Doppler’s
achievement.
In 1886, the Astronomical Institute of the
Charles University in Prague was estab-
lished by August Seydler, who served as
its first director. He elaborated sophisti-
cated methods for determination of orbits
of minor planets and wrote textbooks
on theoretical physics and many popular
articles on astrophysics. In the last de-
cade of the 19th century, Seydler’s suc-
cessor, Gustav Gruss, observed variable
stars and carried out visual observations
of stellar spectra.
servatory at Ondřejov near Prague. In
1928, Josef Jan Frič donated his observ-
atory to the state.
After the establishment of the Czechos-
lovak Academy of Sciences in 1953, the
former State Observatory in Clementi-
num, together with the Ondřejov Observ-
atory, became the basis of the Astro-
nomical Institute of the Czechoslovak
Academy of Sciences. One historical root
of this institution thus reaches to 1722,
when Clementinum observatory was
established by the Jesuits, and another
to the Frič’s foundation of the Ondřejov
Observatory in 1898, which was inspired
by the interest of his poetic godfather
Jan Neruda.
Current Facilities
The Astronomical Institute of the Acad-
emy of Sciences of the Czech Republic
has a prominent position as the largest
professional place in the Czech Republic.
The names of the Institute’s four scien-
tific departments suggest a lot about its
scientific profile: Solar Physics; Interplan-
etary Matter; Stellar Physics; and Galax-
ies and Planetary Systems.
The Solar Physics department focuses
on active solar phenomena such as
sunspots, solar flares and coronal mass
ejections. The data coming from the
instruments of the Ondřejov observatory,
including optical and radio wavebands,
are combined with X-ray and UV obser-
vations from satellites and space probes

Albert Einstein spent one and a half years

(1911–1912) in Prague, being appointed
professor at the German Prague Univer-
sity. During his Prague time he published
eleven papers, including the first one on
the deflection of light by the gravitational
field of the Sun, containing the formula-
tion of the first fundamentals of the theory
of General Relativity.

There is an interesting connection be-

tween astronomy and literature: Jan
Neruda (1834–1891), Czech poet, writer
and journalist, was born in Prague, and
grew up in a colourful part of the town
called Malá Strana (Lesser Quarter). His
first laurels came from the success of
his collection of poetry called “Cosmic
Songs” that documented his good astro-
nomical knowledge. The song No. XXII
“… are there frogs there, too?” was, on the
occasion of the 26th General Assembly
of the IAU in Prague, translated by Robert
Russel into English. Jan Neruda was
godfather of Josef and Jan Frič, the sons
of the revolutionary Josef Václav Frič, who
were born in Parisian exile. Both brothers
shared an interest in astronomy with
their godfather. When they came back to
Bohemia, they started their astronomical
observations, at first with the support
of professor Vojtěch Šafařík. Later, they
established a factory in Prague, which Figure 4: The observation of fireballs has a long
tradition at the Ondřejov Observatory. The photo-
produced optical devices, and developed graph taken on 12 September 1923 by Josef
astronomical instruments. After the death Klepešta shows the M31 galaxy in Andromeda with
of younger brother Jan, Josef built an ob­- a fireball in front.

 The Messenger 128 – June 2007

Figure 5: Ha filtergram (left panel) and correspond-
ing Ha spectrum (right) of a solar active region as
obtained by the Ondřejov large horizontal solar tele-
scope HSFA2 on 4 July 2006. Note the Doppler
shifted features in the spectrum (wavelength increas-
ing to the right).

to have a complex multi-wavelength in-

Photo: J. Havelka, Astronomical Institute, Academy of Sciences of the Czech Republic

formation on these fast non-equilibrium
phenomena. The observations are com-
pared to numerical simulations trying to
discover the basic mechanisms ruling the
Sun. The main goal is the study of mag-
netohydrodynamic and radiative proc-
esses in the solar plasma, which offers
conditions unachievable in laboratories
on Earth. The Sun-Earth relation and
cosmic weather are also studied with a
lot of geophysical and other practical
applications.

The department of Interplanetary Matter

operates the European network of all-sky
cameras, which records meteors, par-
ticularly the brightest of them, which are
called fireballs. This network holds the
world’s primacy in finding meteorites upon
observation of their passage through
the Earth’s atmosphere. In 1959, parts of
the meteorite Příbram were found upon
calculation of its path through the Earth’s
atmosphere based on photographs tak-
en by all-sky cameras. The network of
all-sky fireball cameras and spectral cam- Figure 6: The 2-m
telescope in Ondřejov.
eras has been in continuous operation
since 1963 and is currently the only one
in the world. The atmospheric fragmen- rotation and the detection of near-Earth cially binary stars. The photometric and
tation and structural properties of me- asteroids. In recent years, new discover- spectral observations are combined to
teoroids, chemical composition of me- ies have been made of several binary and disentangle the orbital parameters of bi-
teoroids from spectroscopy, physics of very rapidly rotating asteroids. naries. The use of spectroscopy requires
meteor radiation and ionisation, and long- the development of the theory of stellar
term activity of meteor showers from The Stellar Physics department operates atmospheres, as well as other methods
radar observations, are studied. In 1993 a the largest telescope on the territory of of processing and evaluation of observa-
new programme of CCD observations the Czech Republic, which has a primary tional data. Interacting binaries are usu-
of asteroids using a dedicated 0.65-m mirror of 202 cm diameter. The main ally strong sources of X-ray radiation as
telescope started with the focus on their focus is on the study of variable, espe- observed from space probes.

The Messenger 128 – June 2007 

The Organisation Palouš J. and Hadrava P., Astronomy in the Czech Republic
20

20

20

20

20
10

10

10

10

10
X (kpc)

X (kpc)

X (kpc)

X (kpc)

X (kpc)
0

0
–10

–10

–10

–10

–10
1590 1610 1630 1650 1670

40 50 60 70 20 30 40 50 –10 0 10 20 – 30 – 20 –10 0 – 60 – 50 – 40 – 30
Z (kpc) Z (kpc) Z (kpc) Z (kpc) Z (kpc)

0 10– 5 2 × 10– 5 3 × 10– 5

ρICM
20

20

20

20

20
10

10

10

10

10
X (kpc)

X (kpc)

X (kpc)

X (kpc)

X (kpc)
0

0
–10

–10

–10

–10

–10
1590 1610 1630 1650 1670

40 50 60 70 20 30 40 50 –10 0 10 20 – 30 – 20 –10 0 – 60 – 50 – 40 – 30
Z (kpc) Z (kpc) Z (kpc) Z (kpc) Z (kpc)

–6 –4 –2 0
log ρICM

Figure 7: Gas stripping in galaxy clusters: formation
of a bow shock in front of a galaxy as seen in sim-
ulations using N-body SPH with a gravity tree code.
From Jáhym, P., 2006, Ph.D. Thesis, Charles Uni-
versity Prague, and Université Pierre et Marie Curie,
Paris.
The department of Galaxies and Plane-
tary Systems resides in the Prague part
of the Astronomical Institute, but it has its
own zenith telescope in Ondřejov. It deals
with the study of the rotation of the Earth,
and with the theoretical problems of So-
lar System dynamics and exoplanets. The
trans-Neptunian bodies are studied in-
cluding the resonance in the Kuiper Belt.
Triggered star formation in the turbulent
interstellar medium is studied both in
observations and simulations. The influ-
ence of gas recycling on galaxy evolution
across a Hubble time, including gravi-
tational and hydrodynamic processes, is
considered (see Figure 7 for example).
Study of relativistic astrophysics includes
accretion processes and resonances
in discs surrounding black holes. An ex-
ample is shown in Figure 8 where the
predicted variations of the X-ray spectrum
of an accretion disc at eight successive
phases are plotted during the entire orbit
of a single spot located at a distance of
seven gravitational radii from the black
hole and rotating clockwise. A distant ob-
server is placed at an inclination of 30 de-
grees with respect to the disc normal and
sees the system from the left side. Su-
perposition of several tens of such spots
then contributes to the source variability
as the spots arise, trace some part of the
orbit, and gradually fade away (c.f. Eckart
et al. 2006).
In the Institute of Physics of the Academy
of Sciences of the Czech Republic, a
research group participates in the project
of the Pierre Auger Observatory which
is building an array of telescopes in Ar-
gentina to detect showers of high-energy
cosmic rays. The contribution of the
Czech Republic includes mainly the fabri-
cation of hexagonal mirrors for two the
Schmidt cameras observing ­Cherenkov
radiation in the Earth‘s atmosphere.
The Astronomical Institute of the Charles
University, which is an integral part of the
Faculty of Mathematics and Physics, has
mainly educational duties in teaching
undergraduate and graduate courses of
astronomy and astrophysics. The re-
search areas include the spectroscopic
properties of hot stars, particularly in
binaries and multiple systems, and explo-
ration of the nature of Be stars. Stud-
ies of physical properties, dynamics and
evolution of small bodies of the Solar
System include the Yarkovsky effect and
observations mainly deal with orbits
and light curves of asteroids, including
new discoveries of such bodies. Studies
in cosmology, with special emphasis
on the properties of gamma-ray bursts,
are performed. Studies in history of as-
tronomy focus particularly on the Bohe-
mian region.
The Institute of Theoretical Physics of the
Charles University is also a part of the
Faculty of Mathematics and Physics. Stu-
dents are educated in mathematics and
computing, and in theoretical physics,
including classical and quantum mechan-
ics, elementary particle physics, thermo-
dynamics, relativistic physics and cosmol-
ogy. Research areas include gravitational
collapse with small non-spherical per-
turbations, black-hole electrodynamics,
theory of gravitational radiation, exact
solutions of Einstein’s field equations and
cosmological perturbation theory.
The Institute of Theoretical Physics and
Astrophysics at the Faculty of Sciences,
Masaryk University in Brno, provides ba-
sic courses in astronomy and astrophys-
ics at the undergraduate and graduate
level. Research includes the physics of
hot stars, non-LTE models of stellar at-
mospheres, variable stars, carbon stars,
K-type giants and properties of atmos-

 The Messenger 128 – June 2007

 15 15 15
FE (arbit. units)

FE (arbit. units)

FE (arbit. units)
10 10 10

5 5 5

0 0 0
10 10 10
log E (keV) log E (keV) log E (keV)

15 15
FE (arbit. units)

FE (arbit. units)
10 10

5 5

0 0
10 10
log E (keV) log E (keV)

15 15 15
FE (arbit. units)

FE (arbit. units)
FE (arbit. units)

10 10 10

5 5 5

0 0 0
10 10 10
log E (keV) log E (keV) log E (keV)

Figure 8: The predicted variations of the X-ray spec-

trum of an accretion disc around a galactic nucleus
black hole are plotted. (From Goosmann R. W. et al.,
in “The X-ray Universe 2005”, edited by A. Wilson,
El Escorial, Madrid, Spain, in press.

pheric extinction. The Institute manages
the Masaryk University Observatory at
Kraví hora equipped with a 0.6-m reflec-
tor with a CCD camera.
The Relativistic Astrophysics Group of the
Institute of Physics of the Silesian Univer-
sity in Opava is active in the fields of neu-
tron stars, black holes and cosmology.
The anisotropies of the CBR are explored
as well as L-CDM cosmological models.
The Czech Republic can be proud of its
unique net of planetaria and public ob-
servatories. Their primary goal is to pro-
vide natural science knowledge to a
broad public and to demonstrate the sci-
entific method by special astronomical
research. The Klet’ Observatory, which
is equipped with a 1-m telescope with a
CCD detector, has its research pro-
gramme dedicated to the astrometry of
asteroids with unusual orbits and com-
ets, providing precise determination of
their orbits. The main part of the work is
devoted to the near-Earth orbits. The
data are provided to the Central Bureau
for Astronomy Telegrams of the IAU.
Other observatories include the Nicolaus
Copernicus Observatory and Planetarium
in Brno, Štefánik Observatory and Plane-
tarium in Prague, Rokycany Observatory,
Úpice Observatory, Valašské Meziříčí Ob-
servatory, and others.
This brief review presents the past and
current status of astronomy in the Czech
Republic, which is the starting point for
the entrance of the Czech Republic into
ESO.
References
Eckart A. et al. 2006, The Messenger 125, 2
Hadrava P. 2006, “The European Southern
Observatory and Czech Astronomy”, Academia,
Praha
Palouš J., Vondrák J. and Šolc M. 2002, “Astronomy
and Astrophysics in the Czech Republic”, in
“Organizations and Strategies in Astronomy III”,
ed. A. Heck, Kluwer, 163

The Messenger 128 – June 2007 

 HAWK-I, the new VLT

Telescopes and Instrumentation near-IR camera, as-

sembled and being
tested in the ESO Gar-
ching instrumentation
laboratory.

Photo: H. H. Heyer, ESO

Telescopes and Instrumentation
FORS1 is getting Blue: New Blue Optimised Detectors
and High Throughput Filters
Thomas Szeifert, Roland Reiss, Pedro
Baksai, Sebastian Deiries, Carlo Izzo,
Emmanuel Jehin, Mario Kiekebusch,
Sabine Moehler, Kieran O’Brien,
Emanuela Pompei, Miguel Riquelme,
Gero Rupprecht, Tzu-Chiang Shen
(all ESO)
QE (%)
Ground-based observations in the ul-
traviolet part of the electromagnetic
spectrum are notoriously difficult owing
to absorption of the atmosphere, of
optical elements and the poorer effi-
ciency of detectors. Now that CCD de-
tectors with excellent UV response and
cosmetic quality have become avail-
able, it was time to optimise FORS1 for
imaging, low-resolution spectroscop-
ic and polarimetric observations in the
blue-UV. As a bonus, a new set of
broadband filters with very high trans-
mission and carefully defined filter
bands were installed.
FORS – A brief history
FORS1 and FORS2 are two of the first-
generation VLT instruments that were
built by an external consortium (Landes-
sternwarte Heidelberg, Universitätsstern-
warte München and Universitätsstern-
warte Göttingen, all in Germany). FORS1
was the first of the scientific facility instru-
ments and saw first light at the Casse-
grain focus of VLT-ANTU, on 15 Septem-
ber 1998 (Appenzeller et al. 1998). FORS2
followed in 2000 on VLT-Kueyen. The
FORSes also served on Melipal and
Yepun; currently they are installed again
on Antu (FORS2) and Kueyen (FORS1).
They are amongst the most scientifically
productive instruments of the VLT: more
than 750 refereed papers have been
published to date with both FORSes.
Many of these papers have achieved high
scientific impact factors as indicated by
nearly 20 000 citations.
Soon after it entered regular science op-
erations, FORS2 received a major up-
grade when its original 2k × 2k Tektronix
detector was replaced (effective April
2002) by a mosaic of two red-optimised
MIT/LL CCDs. Further upgrades included
several prototype volume-phased
holographic grisms (VPHG) that greatly
boosted its scientific productivity. We
100
90
80
70
60
50
40
30
20
10
0
300 400 500 600 700
Wavelength (nm)
intended to follow this highly successful
example by adding similar capabilities
to FORS1. The first stage in this process
was the addition of the 1200 B VPHG
that opened new opportunities for stellar
and extragalactic observations by dou-
bling the spectral resolution at very high
grism throughput.
The upgrade
Following the purchase of the VPH grisms
and the success of the new red detector
mosaic of FORS2, it was clear that an
upgrade of FORS1 with a blue-sensitive
new mosaic would be a good comple-
ment to FORS2. The prospect of having
the same detector format for both in-
struments was another strong driver: it al-
lows essentially the same control and
data-reduction software to be used with
FORS1 and FORS2. When the hardware
and the resources to carry out the up-
grade became available, we seized the
opportunity.
The Garching Optical Detector Team
(ODT) prepared the hardware part of the
upgrade, a pair of E2V blue-sensitive
chips. ODT was in particular responsible
for the selection and characterisation
of the CCDs, the mounting and adjust-
ment of the mosaic and the preparation
of the detector control system (FIERA).
The CCDs are named “Marlene”, former-
ly the detector in the UVES blue arm,
and “Norma III”, one of the CCDs from
the batch procured for OmegaCAM.
The choice was motivated by a dramatic
increase in the quantum efficiency of
Figure 1: Comparison
of the CCD efficiencies
of FORS1 and FORS2
before and after the re-
spective upgrades.
Black: Tektronix CCDs;
Red: FORS2 MIT CCDs;
Blue: FORS1 E2V CCDs.
The dotted part indi-
cates the range affected
by fringing (see text).
800 900 1000
these two chips in the blue-UV compared
to that of the existing Tektronix CCD, as
can be seen in Figure 1.
The significant boost in the quantum
efficiency of these CCDs has been
achieved by subjecting them to a treat-
ment of soaking in synthetic air, com-
bined with UV light flashing. The precise
physics of this effect is not known, how-
ever, the effect is reproducible and suf-
ficiently long-lived so that it can be ap-
plied to an instrument in regular science
operations. For a description of the proc-
ess see Baade et al. (2005). Additional
reasons for choosing these new detec-
tors are that they can be read out much
faster and have a lower read-out noise
level, which is crucially important for dark-
time spectroscopy and narrow-band im-
aging in the blue and UV. Finally the cos-
metic quality of these detectors is far
superior to that of the generation of the
Tektronix detector used up to now with
FORS1.
The properties of the new CCD mosaic
are: 4096 by 4096 pixels, 15 µm square,
binned 2 × 2 by default. The pixel scale is
0.25? per (binned) pixel and the inter-chip
gap width is 12.2?. The read-out noise
as measured during the commissioning is
about three electrons.
Strong support came from the Garching
Integration and Cryo-Vacuum Depart-
ment that prepared the cryostat and the
various mechanical pieces required
for the upgrade. An important part of the
upgrade was the adaptation of the var-
ious pieces of software affected by
The Messenger 128 – June 2007 
Telescopes and Instrumentation Szeifert T. et al., FORS1 is getting Blue

the upgrade: Observation Software (OS), Figure 2: Old (left)

meets new (right) – the
templates and the Observer Support
two detector systems
Software (OSS) tool FIMS. These up- side by side on their
grades were done by members of the dedicated carriages
Paranal software department who greatly next to FORS1 in the
enclosure of Kueyen.
benefited from the fact that FORS2 al-
ready uses a similar mosaic and similar
control software. Finally, the Exposure
Time Calculator (ETC) and data-reduction
pipeline had to be adapted and tested.

Installation and Commissioning was a

joint enterprise of theParanal Science
Operation Team and the Garching Instru-
mentation Division (see Figure 2). After
two commissioning runs the upgraded
system was certified in early April 2007
ready for science operations. This is the
reason why the upgrade was only an- 100 Figure 3: Transmission curves of the
old (continuous lines) and new (dotted
nounced in the Call for Proposals (CfP) 90
lines) FORS filters as measured in the
for Period 80. It would however have 80 ESO Optics Lab for Bessel UBVRI and
been hard to justify mothballing an excel- z. The thin and thick red lines indicate
70
lent CCD system for six months and con- the FORS1 R_BESS and the FORS2
Filter response (%)

60 R_SPECIAL filter, respectively.

tinuing operations with an old, inferior
one. So it was decided to go already into 50
P79 with the upgraded system and ad- 40
just the schedule where necessary at 30
short notice in coordination with the PI’s.
20

We also ordered a set of dichroic high 10

throughput filters for the U, B, V and 0

300 400 500 600 700 800 900 1000
g bands to take further advantage of the Wavelength (nm)
new blue sensitivity of the mosaic. We
specified a very high transmission, very 100 Figure 4: Transmission curves of the
old (continuous lines) and new (dotted
carefully chosen central wavelengths and 90
lines) FORS filters as measured in the
full width half maximum values so that 80 ESO Optics Lab for SDSS ugriz filters.
the photometric flux can be nicely trans- The thin and thick red lines indicate
70
ferred to Vega magnitudes with standard the FORS1 R_BESS and the FORS2
Filter response (%)

60 R_SPECIAL filters, respectively. The

stars selected from Landolt (1992). Col-
[O i] line at 557.7 nm falls exactly in the
our corrections are typically smaller than 50
gap between g and R.
0.1 magnitudes. The resulting filter set (to- 40
gether with the already existing R_SPE- 30
CIAL, I_BESS and z_GUNN filters) either
20
matches the Bessel (1990) definition
of the UBVRI or the Sloan Digital Sky Sur- 10

vey (SDSS; Fukugita et al. 1996) ugriz 0

systems. The latter system has achieved
great acceptance in the scientific com-
munity due to the large impact of the
SDSS. Figures 3 and 4 compare the old
FORS1/2 filters with the new ones, the
latter manufactured by Asahi (Japan).
Note that u_HIGH, b_HIGH, v_HIGH and
R_BESS filters are only offered with
FORS1 and the R_SPECIAL filter only
with FORS2.
The new “HIGH” filters are available with
FORS1 since April 2007 in Visitor Mode
300 400 500 600
Wavelength (nm)
only, because they were not yet ready
and characterised when the CfP for Pe-
riod 79 went out in September 2006.
Let the game begin …
The new detector system saw its first sky
light on 30 January 2007. During a first
commissioning run lasting three nights
700 800 900 1000
we performed a thorough characterisa-
tion of the quantum efficiency, not only of
the detectors but also of the new filters.
Observations of numerous photometric
and spectrophotometric standard stars
proved that our expectations, based
on the laboratory data, were correct. The
new detector alone gave an improve-
ment of 0.8 and 0.4 magnitudes in U and
B, respectively, and the performance

10 The Messenger 128 – June 2007

dropped (as expected) only slightly in the
I-band. Using, in addition, the new high
throughput filters resulted in a spectacu-
lar gain of 1.3 mag. in U and 0.8 and
0.3 mag. in B and V.
The response given here has been calcu-
lated from the photometric zero points
measured during the first commissioning
run, according to the following strategy.
First we calculated the Vega flux, integrat-
ed over the filter curves. From the Vega
flux we calculated the zero points for
100 % instrument and telescope through-
put in magnitudes (27.41, 29.12, 29.21,
29.18, 28.78 in UBVRI and 27.93, 29.47,
29.33 and 29.90 in the ubvg filters, re-
spectively) for incoming photons/sec at
the 8-m aperture of the VLT (see Table 1).
The overall instrument response can then
be easily derived from the measured
photometric zero points at zero airmass.
Similarly the Vega zero points at 100 %
response were calculated for the VIMOS
UBVRI filters (27.93, 29.37, 28.96, 29.05
and 28.91 mag.) and for the FORS2
R_SPECIAL filter (29.33 mag.). The res-
ponse is then given in units of detected
electrons per incoming photon including
the telescope, the FORS longitudinal
Atmospheric Dispersion Corrector (ADC),
the instrument optics and detector res-
ponse, but not the filter transmission.
This response is given in parentheses in
Table 1 and demonstrates the high per-
formance of FORS1, FORS2 and VIMOS
in all filters.
In addition the new g-band filter opens a
new observation window. It collects the
flux from the astronomical targets over
a wide wavelength range were the night
sky is very dark and the atmospheric
transmission is high (390 nm to 550 nm).
A second commissioning run (five nights
in March/April 2007) finally verified the
functioning of the system in all supported
observing modes (imaging, long-slit
and multi-object spectroscopy, imaging
and spectropolarimetry).
… even with some adverse effects
There is however a significant price to
pay for high quantum efficiency in the
blue-UV: the fringe pattern at near-infra-
red wavelengths, beyond approxima-
Filter FORS1 FORS1 FORS2 VIMOS Table 1: Zero points and instrument response for
TEK e2v MIT e2v FORS1, FORS2 and VIMOS. Note that the VIMOS
U 25.20 26.53 n/a 26.5 U-filter red cut-off is at 395 nm while it is at 385 nm
for FORS1. Similar differences exist also for the
(0.13) (0.28) (0.27)
other filters; therefore zero points do not allow a di-
B 27.70 28.48 27.70 28.20
(0.27) (0.40) (0.27) (0.34)
rect comparison of the absolute instrument res-
ponse. The instrument response in detected elec-
g n/a 28.89 n/a n/a trons per incoming photon after eliminating the
(0.39) effect of the different filter sets is given in parenthe-
V 28.05 28.33 28.10 27.90 ses.
(0.34) (0.39) (0.39) (0.38)
R 28.00 27.96 28.40 27.90
(0.34) (0.33) (0.42) (0.35)
I 27.15 26.99 27.70 27.0
(0.22) (0.19) (0.37) (0.17)
tely 700 nm, increases significantly as pixel has the same wavelength for targets,
compared to the old (somewhat thicker) sky and flat field lamps. The correction
Tektronix detector and the spatial fre- should therefore be possible. In practice,
quency of the fringes also increases however, the slightly different light path
strongly. The fringing has a heavy impact of the calibration light as compared to the
on high signal-to-noise observations. telescope optical path, together with
the small flexure of the FORS instrument,
For imaging observations the fringes will prevents perfect fringe correction using
remain at high amplitude in the sky back- flat field spectra. In first tests, we ob-
ground, on account of its OH emission- tained signal-to-noise ratios of up to 15 at
line spectrum, after flat fielding with the wavelengths greater than 700 nm (see
solar spectrum twilight flat fields. The sky Figure 5).
background with fringes is best sub-
tracted by obtaining an averaged sky im- Many scientific projects with FORS1 are
age cleaned from astronomical sources, focused on detecting extremely faint
scaled to the sky level of the individu- objects at very low signal-to-noise ratios
al exposures. More commonly observers or will concentrate on the shorter wave-
obtain “super flats” from the night sky length range after the blue optimisation.
airglow images observed in a jitter se- To demonstrate the performance of
quence. While this method leads to a FORS1, Figure 6 shows a spectrum of a
more pleasing image with a flat sky back- z = 2.42 quasar of g = 20.4 magnitudes
ground, it strongly compromises the pho- which was obtained in only 15 minutes of
tometric accuracy, which better matches integration time.
a continuous or solar-type spectrum than
the OH airglow spectrum. In both imaging and spectroscopy it is
mandatory to apply jitter and nodding
In spectroscopy, flat fields obtained with techniques to obtain good sky subtrac-
continuous light calibration lamps can, tion. The fringes however will not be cor-
at least partly, correct the fringes. In the- rected in the extracted spectra of science
ory the light that is detected by every targets and standard stars by the nod-
Figure 5: A demonstration of fringe correction with
the new FORS1 detector. The top panel shows part
of a spectroscopic (master) screen flat. The next
two panels show parts of two bias-subtracted sec-
tions from 1200 s spectra taken with an offset along
the slit (positions A (left), and B (right)). The two
panels below show the same parts of the spectra,
now divided by the master flat (top). Fringes are
only partly corrected by the continuous light screen
flat fields. The difference image of these two expo-
sures (bottom) still contains significant sky back-
ground flux. A better result could be achieved using
a sequence of four exposures (A-B-B-A).
The Messenger 128 – June 2007 11
Telescopes and Instrumentation
ding technique. More details on this topic
will be available on the FORS web pages.
It should be noted that FORS2 has ex-
traordinarily low fringe amplitude with its
red-optimised MIT detector combined
with the very low flexure it shares with its
twin, FORS1. Users with strong require-
Flux (e –/sec)
ments on fringe correction are therefore
encouraged to choose FORS2.
In summary, the successful blue up-
grade of FORS1 leads to a very promis-
ing complement to FORS2 that will fur-
ther enhance the scientific productivity of
this efficient and reliable pair of instru-
ments.
Acknowledgements
We are grateful to: the FORS Instrument Opera-
tions Team (ESO) for continuous support through all
stages of the upgrade project: Magda Arnaboldi,
Paul Lynam, Palle Møller, Ferdinando Patat, Marina
Rejkuba; the PAO Engineering Department for effi-
cient support during the commissioning, in particu-
lar Roberto Castillo and Pascal Robert; and finally
Walter Seifert (Landessternwarte Heidelberg) for his
support in the definition of the new filter set.
12 The Messenger 128 – June 2007
Szeifert T. et al., FORS1 is getting Blue
6
SDSS J090847.08+010114.1
4
2
0
400 500 600 700 800
Wavelength (nm)
References Figure 6: Extracted, airmass-corrected spectrum of
the quasar SDSS J090847.08+010114.1 (20.4 mag in
Appenzeller I. et al. 1998, The Messenger 94, 1 g-band, z = 2.42). Note that the spectrum (exposed
Baade D. and the Optical Detector Team 2005, in for 15 minutes during dark time without order sepa-
“Scientific Detectors for Astronomy”, eds. J. E. ration filter) is also shown in the red and near-infrared
Beletic, J. W. Beletic and P. Amico, Astrophysics spectral range where the fringes are strongest. The
and Space Science Library, Springer, 73 strongest three emission lines (from left to right) are
Bessel M. S. 1990, PASP 102, 1181 the hydrogen Lyman alpha line, C iv and C iii] lines.
Fukugita M. et al. 1996, AJ 111, 1748
Landolt A. U. 1992, AJ 104, 372
VLT FORS1 image of the bubble nebula N76 around
the hot binary star AB7 in the Small Magellanic
Cloud, based on three exposures through narrow-
band filters isolating doubly ionised helium (He ii, in
blue), doubly ionised oxygen ([O iii], in green) and
singly ionised hydrogen (H-alpha, in red). The image
measures 400 by 400 arcseconds and north is up
and east to the left. The binary system AB7 (the
bright stellar image in the centre of the nebula) con-
sists of one evolved massive Wolf-Rayet star and a
companion O-type star. The very high temperature of
the stars is responsible for the centred He ii nebula
(blue region enclosed within the yellower ring). To
the north-east, just outside the nebula, a small net-
work of green filaments is visible, a remnant of an
earlier supernova explosion. See ESO PR 08/03 for
more details.
Telescopes and Instrumentation
Towards Precision Photometry with FORS:
A Status Report
Wolfram Freudling, Palle Møller,
Ferdinando Patat, Sabine Moehler,
Martino Romaniello, Emmanuël Jehin,
Kieran O’Brien, Carlo Izzo, ­
Eric ­Depagne, Emanuela Pompei,
­Dominique Naef, Gero Rupprecht,
Arto Järvinen (all ESO)
The two FOcal Reducer and low-dis-
persion Spectrographs (FORS) are
the primary imaging instruments for the
VLT. Because they are not direct
imaging instruments, the accuracy of
photometry which can routinely be
obtained is limited by significant sky
concentration and other effects. This
article reports on the progress of
a long-term project to improve the
photometric calibration of the FORS
instruments.
The calibration plan for the FORS instru-
ments calls for observations of photo-
meric standards in each clear night. The
primary purpose of these observations
is to monitor instrument performance.
The same data are also used to calibrate
science observations for programmes
where 5 to 10 % photometric accuracy is
sufficient. Two years ago we started the
FORS Absolute Photometry (FAP) project
to characterise the photometric perform-
ance of the FORS instruments and inves-
tigate if and how the routine calibration of
the instruments can be modified to offer
improved photometric zero points (ZPs).
An additional goal of this programme is to
develop procedures to allow users to get
more accurate photometric calibration.
The results are available as two internal
ESO reports, Møller et al. (2005) and
Freudling et al. (2006). In this article, we
present a brief overview of the issues re-
lated to photometry with the FORS in-
struments, and the current status of FAP.
Our work has so far concentrated on
the FORS1 camera, but the most of the
findings will equally apply for FORS2.
Relative photometry with FORS1
Accurate photometry starts with reliable
and stable flatfielding. We used the
huge collection of twilight flats in the VLT
archive to investigate the quality of the
flatfields commonly used for the reduc-
tion of FORS images. In order to investi-
gate the stability of the flatfields, we
computed the mean of bias-subtracted
flatfields, and divided each individual
frame by this mean. This removed the
stable part of the flatfields. Visual inspec-
tion revealed that the structure in the
flatfields consists of a temporally constant
pattern, superimposed on large-scale
fluctuations which rapidly change in time.
The contrast of the constant pattern
is higher in bluer bands. Interestingly, we
found a correlation of some of the pat-
terns with the adaptor rotator angle. The
FORS instruments are mounted on
adapter rotators which compensate for
the sky-field rotation inherent to the VLT
alt-azimuth mounting. Part of the struc-
ture in the flat field rotates rigidly with
the angle of the rotator. This is illustrated
in Figure 1. This pattern in the flatfield
must be external to FORS1 and might be
due to reflections and/or asymmetric
vignetting within the telescope or the
adapter itself.
A high signal-to-noise version of the ro-
tating structure is shown in Figure 2. This
image was created by counter-rotating
B flatfields by an amount equal to the ro-
tator angle and then computing the me-
dian of the rotated flatfields. If there was
no correlation between the structure
Figure 1: A sequence of B FORS1 sky-flats divided
by the median of all flatfields. The intensity scale
range is 3 %. The rotator adapter angles (−105, −73,
− 35, 0, + 30, and +70) are indicated by a blue arrow
in the upper right corner.
The Messenger 128 – June 2007 13
Telescopes and Instrumentation
in the flatfields and the rotator angle, then
the structure of the individual flatfields
should average out and the median would
be smooth and flat. Instead the opposite
can be seen in Figure 2. A finger-like pat-
tern, which is already visible in the indi-
vidual flats shown in Figure 1, stands out
with increased signal-to-noise. This dem-
onstrates that this is a rotating feature.
The peak-to-peak amplitude of the pat-
tern in the median frame is about 1 %. In-
spection of individual images in the stack
shows that the amplitude varies substan-
tially among the individual flatfields.
Twilight flats, as routinely obtained each
night, differ from each other by as much
as 5 %. If such flatfields are applied to
science data, the relative photometric ac-
curacy is limited to about 5 %. Even when
controlled for rotator angle, flatfields dif-
fer from each other by an amount which
questions the feasibility of per-cent-level
photometric accuracy with FORS1. A key
question is whether these fluctuations
reflect true differences in the end-to-end
throughput of FORS1. In that case, rel-
ative and therefore absolute accuracy at
the per cent level simply cannot be ob-
tained with FORS1. A more likely explana-
tion is that the flatfields are flawed and do
not represent the throughput of FORS1.
In that case, the task is to find the true
flatfield which should be applied to data
so that the photometric ZPs are constant
over the whole detector.
We experimented with different proce-
dures to remove the large-scale pattern
from the raw observed flatfields. For the
investigations described in the follow-
ing sections, we used new observations
to test the quality of the flatfields con-
structed in this manner, and compared
them to the regular “master flats” pro-
duced by combining the routine twilight
flats for that night.
The large variation in the flatfields was
the motivation to take a closer look at
possible variations of the ZP magnitudes
across the detector when using the mas-
ter flat. The goal was to derive a correc-
tion for the master flats and to find a
quantitative estimate of the accuracy of
the finally adopted flatfield. For that pur-
pose, we observed standard fields with a
25-point dither pattern which placed
one relatively bright star on a grid of po-
14 The Messenger 128 – June 2007
Freudling W. et al., Towards Precision Photometry with FORS: A Status Report
Figure 2: A stack of all B sky flats after applying a
rotation around the geometrical centre with an
amplitude equal to the adapter rotator angle. The
intensity scale range is 1%.
sitions across the CCD. This approach
is often nicknamed the “1000 points of
light” approach. In addition, we took im-
ages of the same fields rotating the whole
frame.
Ideal data sets for calibrating images with
suspected ZP variations across the de-
tectors are fields which contain suitable
photometric standards over the whole
dithered region. We therefore selected re-
gions within Stetson fields (Stetson, 2000,
2006) as our targets, which contain a
large number of photometric standards
suitable for 8-m-class telescopes. These
fields proved to be useful for our pur-
pose, and we will investigate whether all
current FORS calibration fields can be
replaced with appropriately selected point-
ings within the Stetson calibration fields.
The simplest and most direct way to in-
vestigate relative ZP changes with
dithered data is to compare the relative
instrumental magnitudes of individu-
al standard stars which are observed
at positions all across the detectors.
A much more sensitive method is to use
the measured magnitudes of any star
which has been observed at two or more
different dither positions. Each individ-
ual star might only provide relative ZP
shifts for a few positions on the detector.
By combining the information from many
stars, the ZP variations over the whole
detector can be reconstructed. We used
singular value decomposition (SVD) to
construct such an image of the ZP varia-
tions from our dithered observations. The
details of the formalism are described
in Freudling et al. (2007). This image can
then be used to correct the flatfield, or
be applied as a second-order flatfielding
step to the science data. The flatfield cor-
rection frame for the R filter is shown
in Figure 3. The peak-to-peak flatfielding
error at the position of the observed stars
is about 30 mmag. It should be empha-
sised that the amplitude of the correction
frame depends on the flatfield used to
process the imaging data, and that the
correction frame therefore only applies to
that particular flatfield.
We tried several methods to remove
the fluctuating and rotating features,
described above, from the master flats
before computing the correction frame.
The most successful approach was to
remove the large-scale pattern by divid-
ing each flat-field by a smoothed version
of itself. This procedure lowers the am-
plitude of the necessary correction to the
flat-fields derived from observations of
standard stars, and therefore improves
the overall flat-fielding.
So far we have derived flatfield correc-
tions for three data sets taken more than
two years apart. The overall shape and
amplitude of the correction frames were
stable over that period. We therefore
­conclude that the best flatfielding with
FORS1 can be obtained by: (1) remov-
ing large-scale features from the master
flats; (2) applying a correction frame de-
rived from observations of standard
fields.
Absolute photometry with FORS1
Nightly ZPs for FORS instruments have
so far been computed assuming a con-
stant extinction for each night. The ex-
tinction, however, varies substantially
from night to night, even when the nights
are photometric. Therefore, ZPs derived
using a mean extinction depend on the
airmass of the measured standard field
and are not useful for accurate photom-
etry. The true instrumental photometric
ZP above the atmosphere, as derived
from extrapolation of the extinction curves,
probably varies much more slowly than
the night-to-night variations of the extinc-
tion. Therefore, better photometric ac-
curacy can be obtained by fitting the ex- Figure 3: R-band flatfield
correction frame.
tinction coefficients for several nights
simultaneously, assuming that the instru-
mental ZP does not change. Such a
scheme can be included in the computa-
tion of the flatfielding correction, so that
a photometric solution is simultaneously
obtained with the relative flatfield correc-
tion.

The measured magnitudes of repeatedly

observed stars can also be used to esti-
mate short-term fluctuations in the extinc-
tion during the night. Each of the more
than 1000 individual stars observed with
our dither pattern can be used for this
purpose. When the weighted average of
all stars is used, changes in the extinc-
tion can be measured with an accuracy
of about 1 mmag. An estimate of fluctua-
tions in the extinction is an automatic by-
product of the fitting procedure described Figure 4: Upper panel: Residuals from
40 the extinction solution as a function
above.
of the estimated magnitude error. The
magnitude error includes the mea-
Residuals (mmag)

One of the tools to independently assess
the quality of the night is the “VLT As-
tronomical Site Monitor” (ASM) which can
be accessed at http://archive.eso.org/
asm/ambient-server. We found that,
at least in the one photometric night used
for FAP, the measured scatter of about
6.8 mmag is very similar to the flux rms
20
surement error from the current obser-
0 vations as well as estimated uncer-
tainties for the magnitudes of standard
stars. Lower panel: The variance of
– 20
the residuals as a function of magni-
tude error. The superimposed line is
– 40
an error model, which includes statisti-
cal errors as well as extinction fluctua-
tions.
30

measured by the ASM monitor. It is

Variance (mmag)

tempting to conclude that the rms from

the ASM can be used as a proxy for
expected rms fluctuations of the ZP. We
plan to further investigate the useful-
ness of the ASM to judge the photometric
quality of the night.
Understanding the achievable accuracy
of photometry requires understanding
the total error budget. In particular, it is
important to understand the accuracy of
our procedures for flatfielding. We used
the residuals of the extinction solution for
a detailed investigation of the error esti-
mates.
Measurement errors were computed for
each measured magnitude from the
read-out and Poisson noise, both for the
pixels used to compute the stellar flux
and for those used to estimate the local
background. The error estimates range
from 2 to 30 mmag. Errors in the listed
standard magnitudes were taken from
Stetson (2000) and Stetson (2006), which
are based on repeated observations in
20
10
0
0 5 10 15
σm (mmag)
different nights. Above, we estimated that
extinction fluctuations during the photo-
metric night for this programme were
≈ 6.8 mmag. In Figure 4 we demonstrate
that extinction variations, statistical errors
and errors in the standard magnitudes
account for most of the residuals of our
photometric solution. A detailed analysis
suggests an upper limit on residual flat-
fielding and other sources of errors of
about 3 mmag. This demonstrates that
precision photometry with per-cent-level
accuracy is indeed possible with FORS1.
One of the important goals of this project
is to define a set of guidelines on how
to achieve a per-cent-level photometric
20
accuracy. The uncertainty in the photo-
metric ZP is obviously an important factor
which determines the final accuracy of
the magnitudes. The FAP observations
contain a large number of standard stars
on each individual image, and the num-
ber of calibration images is much larger
than the number realistically taken for
the calibration of normal science obser-
vations. An important part of the photo-
metric guidelines will therefore be to
find the necessary minimum number of
standard fields needed to achieve a cer-
tain photometric accuracy.
FAP imaged four different Stetson fields
in a single photometric night, which were

The Messenger 128 – June 2007 15

Telescopes and Instrumentation
used to derive a photometric extinc-
tion solution. The magnitude and colour
range, and the consistency of derived
ZPs is similar for all fields. To estimate the
errors on the ZPs from sets of only two
standard field observations, we recom-
puted the ZPs from subsets of the FAP
data. We used every combination of two
standard fields which were taken with
a difference in airmass of at least 0.7.
The distribution of the resulting ZPs is
shown in Figure 5. The distribution has
an almost Gaussian peak but also a long
non-Gaussian tail. In about 10 % of all
cases, the errors on the resulting ZPs is
larger than 3 %. This shows that the
observation of only two standard fields is
insufficient to photometrically calibrate a
night to per-cent-level accuracy.
We then repeated the experiment using
three standard fields. At most one of
the fields in each set was at an airmass
less than 1.3, and the differences be-
tween minimum and maximum airmass
was larger than 0.7. The resulting distri-
bution of ZPs is plotted in Figure 5. Also
shown is a Gaussian with the same
mean, standard deviation and normalisa-
tion as the ZP distribution. It can be seen
that the distribution resembles closely
a Gaussian with a standard deviation of
11 mmag. In contrast to the previous
experiment with only two standard fields,
all ZP errors are less than 3 %. This result
strongly suggests that the use of three
photometric standard fields, chosen with
30
20
N
10
0
27.4 27.5
mz
16 The Messenger 128 – June 2007
Freudling W. et al., Towards Precision Photometry with FORS: A Status Report
the strategy outlined above, leads to an
accuracy for the magnitude ZP of about
10 mmag.
The scatter in the ZPs also suggests that
errors are almost Gaussian when three
different standard fields are used. The er-
ror budget discussed above implies that
the dominant error on the mean magni-
tude of all stars in any of the standard
fields are fluctuations in the extinction, if
the number of standard stars in each
field is large enough and the exposures
sample the airmass between 1 and 2 uni-
formly. For a typical magnitude uncer-
tainty of 10 mmag, about 30 or more
standard stars per field are needed. This
is one of the reasons to use the Stetson
standard fields as opposed to fields with
fewer stars with known magnitudes.
The past and future FORS1 photometric
systems
In April 2007 FORS1 was upgraded with
a new CCD mosaic and a new set of
broad-band filters (see article on page 9).
During February and March 2007 we
therefore carried out an intensified ver-
sion of our usual photometric calibration
plan, where we employed all the tools
and methods developed during the FAP.
This data set includes dithers and rota-
tions which makes it possible to derive
flatfield corrections directly from the cali-
bration data. The calibrations obtained
Figure 5: Distributions of zero points
determined from two and three stand-
ard observations (red and blue histo-
grams respectively). The dashed line is
a Gaussian with a s of 11 mmag.
during those two months will therefore
serve to determine the best possible, and
final, photometric characterisation of the
now retired system which has been used
since the start of VLT operation.
This “Final Effort” on the retired FORS1
photometric system will also serve as a
model for the design of our future FORS
calibration plan. Our pilot programme
has shown that per-cent-level photometry
with FORS is indeed feasible. The key
requirement for this is the ability to com-
pute the flatfield corrections. In order
to make this possible for a large number
of programmes, those corrections must
be determined directly from the calibra-
tion plan data, which in turn places more
demands on our calibration plan. In par-
ticular there needs to be a large number
of moderately faint standards in each field.
We are currently investigating whether
appropriate regions for this purpose can
be found within the Stetson standard
fields. Using these fields will therefore
require that we measure U-Band magni-
tudes for the standards in those fields.
(Stetson, 2006).
FAP has shown that it is possible to
achieve per-cent-level photometry with
FORS1 with moderate effort. Over
the next year, we plan to prepare new
photometric standard fields suitable
to obtain more accurate photometric
solutions and simultaneously derive
flatfield corrections. If this programme
proves to be successful, we plan to
incorporate new flatfielding algorithms
into the FORS pipeline.
References
Freudling W. et al. 2006, “The FORS Absolute
Photometry Project”, VLT-TRE-ESO-13100-4006
Freudling W. et al. 2007, in “The Future of Photo-
metric, Spectrophotometric, and Polarimetric
Standardization”, ed. C. Sterken, ASP Conference
Series 364, 113
Møller P. et al. 2005, “FORS: An assessment of ob-
tainable photometric accuracy and outline of
strategy for improvement (FORS IOT Secondary
Standards Working Group)”, VLT-TRE-ESO-13100-
3808
Stetson P. B. 2000, PASP 112, 925
Stetson P. B. 2006, fields listed at http://cadcwww.
dao.nrc.ca/cadcbin/wdbi.cgi/astrocat/stetson/
query
Telescopes and Instrumentation

Exploring the Near-Infrared at High Spatial and Spectral

Resolution: First Results from CRIRES Science Verification

Ralf Siebenmorgen 1 The VLT cryogenic high-resolution infra- be used to increase both spatial resolu-
Alain Smette 1 red echelle spectrograph CRIRES offers tion and signal-to-noise ratio.
Hans Ulrich Käufl 1 high spatial, spectral and temporal
Andreas Seifahrt 1 resolution spectroscopy from 1 to 5 μm. The scientific potential of the CRIRES
Stefan Uttenthaler 2 Highlights from among the 29 pilot instrument is demonstrated by the many
Arjan Bik1 studies of the CRIRES science verifica- results obtained during three science
Mark Casali 1 tion (SV) runs are summarised. verification observing campaigns which
Swetlana Hubrig 1 have been performed in August, October
Yves Jung 1 2006 and in February 2007. Twenty-nine
Florian Kerber 1 The VLT cryogenic high-resolution infra- pilot studies were granted observing time
Jorge Melnick 1 red echelle spectrograph CRIRES (Käufl totaling 20 000 s integration time. The
Alan Moorwood 1 et al. 2004) is located at the Nasmyth principal investigators (PI) and project ti-
Jean-François Pirard 1 focus A of UT1 (Antu). It provides a resolv- tles of the successfully executed SV
Hugues Sana1 ing power of up to 100 000 in the spec- programmes are given in Table 1. Individ-
Elena Valenti 1 tral range from 1 to 5 μm. CRIRES can ual projects range from studies of the
Lowell Tacconi-Garman1 boost all scientific applications aiming at Earth’s atmosphere, disc structure around
Michael Hilker 1 fainter objects, higher spatial resolution young stars, brown dwarfs (BD), extra-
Francesca Primas1 (for extended sources), spectral and solar planets to pulsation and wind prop-
Pedro J. Amado 3 temporal resolution. Spectral coverage is erties of massive stars, asymptotic giant
Andrés Carmona1 maximised through a mosaic of four branch stars (AGB), structure of the
Ewine F. van Dishoeck 4 Aladdin III InSb arrays providing an effec- Galaxy and astroseismology. In this arti-
Cédric Foellmi 1 tive 4096 × 512 detector array in the focal cle some of the highlights are presented
Miwa Goto 5 plane. A MACAO (Multi-Applications in a sequence beginning with young
Roland Gredel 5 Curvature Adaptive Optics) system can stars, brown dwarf stars and planets,
Eike Günther 6
Bengt Gustaffson 7
Don Kurtz 8 Tabel 1: CRIRES science verification runs: principal investigator (PI)
and project title is given for each programme.
Christopher Lidman1
Hendrik Linz 5
Fabrice Martins 9 ID PI Title
Karl Menten 10 1 Amado NIR spectroscopy of pulsating stars?
Claire Moutou 11 2 Bik Circumstellar discs around massive young stellar objects
Poul E. Nissen 12 3 Carmona Probing the gas in the inner 50 AU of proto-planetary discs
Dieter Nürnberger 1 4 Foellmi Distances to late-type stars
Ansgar Reiners 1 5 Goto H+3 and CO observation toward Superantennae
6 Günther CRIRES for high precision RV of late-type stars
7 Gustaffson CNO abundances in Bulge giants
1
ESO 8 Käufl Search for OH in the disc around HD 163296
2
Institut für Astronomie, University 9 Kerber Determining the atmospheric precipitable water vapour content
Vienna, Austria 10 Kjaer High resolution infrared spectrum of SN 1987A
3
Gr-IAA, Instituto de Astrofísica de 11 Kurtz High time-resolution precision radial velocities of peculiar A stars
Andalucía, Granada, Spain 12 Lidman A high-resolution spectral atlas of the night sky
4
Universiteit Leiden, the Netherlands 13 Linz Disc winds and envelopes associated with BN-type objects
5
Max-Planck Institute for Astronomy, 14 Martins The most massive stars in the GC: binarity and metallicity
Heidelberg, Germany 15 Melnick and Gredel Molecular hydrogen in Doradus 30
6
Thüringer Landessternwarte (TLS), 16 Moutou Transmission spectroscopy of transiting extrasolar planets
Tautenburg, Germany 17 Nissen The abundance of sulfur in metal-poor stars
7
Uppsala University, Sweden 18 Nürnberger Weighing a high mass protostellar candidate
8
University of Central Lancashire, 19 Reiners FeH spectroscopy in ultra-cool dwarfs – CRIRES or UVES?
Preston, United Kingdom 20 Sana Can IR solve the wind clumping question?
9
Max-Planck Institute for Extraterrestrial 21 Seifahrt Effective temperatures and gravities of low-mass stars and BD
Physics, Garching, Germany 22 Siebenmorgen Astrochemistry in dust formation regions
10
Max-Planck Institute for Astronomy, 23 Siebenmorgen Roadmap of astrochemistry during massive star formation
Bonn, Germany 24 Siebenmorgen Bound on time dependency of the fine structure constant
11
L aboratoire d’Astrophysique Marseille 25 Smette Na i D and Ca ii in a z ~ 2 damped Lya system
(LAM), France 26 van Dishoeck CO emission from transitional protoplanetary discs
12
University of Aarhus, Denmark 27 Uttenthaler The C/O and C12/C13 ratios and three dredge-up in bulge AGB stars
13
Hamburger Sternwarte, Universität 28 Uttenthaler Titanium Oxide band heads in the J-band
Hamburg, Germany 29 Valenti Chemical composition of evolved populations in LMC and NGC 1866

The Messenger 128 – June 2007 17

Telescopes and Instrumentation Siebenmorgen R. et al., First Results from CRIRES Science Verification

followed by massive and Asymptotic Gi- 1.0 Figure 1: 12 CO profile of the massive young stellar
object IRAS 16164-5046. From top to bottom, lines
ant Branch stars to targets outside the

Normalised flux
0.8
R(7), R(4) and R(0) are plotted. The R(0) line is more
Galaxy. Finally CRIRES calibration issues 0.6 sensitive to cold material than the other two. Two
are mentioned. 0.4 components can be identified, one belonging to the
12
CO R(7) cold envelope and the other component caused by
0.2 4.6024 micron a wind or outflow (Bik et al.).
0.0
Discs around massive young stellar 1.0

objects and extrasolar planets

Normalised flux

0.8

0.6

The known circumstellar discs around 0.4

massive young stellar objects T Cha and 0.2

12
CO R(4)
4.6254 micron
IRAS 16164–5046 were studied by van 0.0
Dishoeck [ID 26] and by Bik [ID 2], respec- 1.0

tively. The CO first-overtone emission in

Normalised flux

0.8

the K-band probes high-temperature gas 0.6

(T = 3 000 K) and the velocity profile of 0.4

CO R(0)
IRAS 16164–5046 suggests that the hot
12
0.2 4.6575 micron
gas is located in a rotating disc within 0.0
– 20 –10 0 10 20
10 AU of the central star. The CO-funda- Velocity (km/s)

mental transitions in the M-band probe

lower-temperature gas (T = 50–500 K)
and therefore different regions in the cir-
cumstellar material. Figure 1 shows three
transitions of the 12 CO spectrum of IRAS
16164–5046. The R(7) line is more sen-
sitive to higher temperature compared to RV (km/s) RV (km/s) Figure 2: Line flux and RV of the Bra
– 300– 200 –100 0 100 200 300 – 200 –100 0 100 200 and Pfg (top) and various CO iso-
R(0). The spectra reveal two absorption
topes (bottom) of the protostar W33a
components; a saturated blue compo- (Siebenmorgen et al.).
nent which could be caused by an out-
flow or wind, and a red, unsaturated com-
ponent which changes in intensity from Brα Pfγ
the higher to the lower energy level. This
Normalised flux

extra red absorption could be a cold en-

velope surrounding this object.

A full L- and M-band (3 600–5 100 nm)

wavelength scan of the spectrum of the
extremely red and massive protostar
W33A was performed, aiming to achieve
a first road map of the astrochemistry
occurring during the formation process of
4 050 4 052 4 054 4 056 3 738 3 739 3 740 3 741 3 742 3 743
massive stars. A section of it, correspond- Wavelength (nm)
ing to the spectral region near 4 200 nm,
is shown by Käufl et al (2007). Figure 2
compares the rich CO line system (near 12
C16 OJ = 1–0 P7 12
C16 OJ = 1–0 P8

4 730 nm) to the broad Bra and Pfg lines. 13

C16 OJ = 1–0 R5 13
C16 OJ = 1–0 R4 13
C16 OJ = 1–0 R3
Derived radial velocities (RV) for CO,
Bra and Pfg lines are ~ 31 km/s confirm- C OJ = 1–0 R6 C OJ = 1–0 R5 C18 OJ = 1–0 R4
12 18 12 18 12

ing earlier estimates (Roueff et al. 2006). 12

C17OJ = 1–0 P2
Normalised flux

12
C17OJ = 1–0 P1

Carmona [ID 3] obtained K-band spectra

of the classical T Tauri star LkHa 264.
They confirm the previous discovery of the
ro-vibrational n = 1−0 S(1) H2 line. In addi-
tion, thanks to the enhanced sensitivity
of CRIRES, they detect at 2 223.3 nm the
n = 1−0 S(0) H2 transition. However, the
spectra do not reveal the n = 2−1 S(1) 4 725 4 730 4 735 4 740 4 745
H2 line. The line ratios of 1–0 S(0)/1–0 S(1) Wavelength (nm)

18 The Messenger 128 – June 2007

= 0.33 ± 0.1 and 2–1 S(1)/1–0 S(1) < 0.2
indicate that the H2 emitting gas is at tem-
perature below 2 000 K and most likely
thermally excited by UV photons. Both
Normalised flux – continuum
detected lines have a FWHM of ~ 20 km/s
and are spatially unresolved. The mean
FWHM of the PSF in the continuum is
~ 0.36? in the H2 1–0 S(1) spectrum, con-
straining the H2 emitting region to the in-
ner 50 AU of the disc at an assumed
distance of 300 pc. The n = 1−0 S(1) and
the n = 1−0 S(0) H2 lines in LkHa 264 are
single peaked. Modelling of the n = 1−0
S(1) line shape indicates that the disc
has a relatively small inclination. The best
model fit suggests that the disc sur-
rounding LkHa 264 is inclined by 20˚ rela-
tive to the line of sight (Figure 3).
Late-type objects of spectral type M and
later exhibit extremely interesting ab-
sorption bands. In particular, the molecu-
Normalised flux (F λ + constant)
lar band of FeH shows sharp absorption
lines that can be used for precision spec-
troscopy, a rare feature for late-type stars
and brown dwarfs. These lines appear
in the spectral range close to ≈ 1000 nm,
at the red end of the UVES spectral cov-
erage and at the blue one of CRIRES. A
performance test of both instruments
in this overlapping region was carried out
by Reiners [ID 19] on GJ 2005A, a M5.5
star, observed with CRIRES and MACAO,
while Gl 406, a star of the same spectral
type, was observed with UVES. The FeH
absorption of both stars resembles each
other very closely. The UVES data were
taken at a resolution of 50 000, and for
comparison CRIRES data were rebinned
to match this lower resolution. The UVES
exposure of Gl 406 (V = 13.54) yielded
a SNR of 60 after 200s whereas the
CRIRES spectrum on the much fainter
star GJ 2005 A (V = 15.42) gives a SNR
of 50 after 480 s. Therefore, in this over-
lapping wavelength region, UVES re-
quires roughly twice the observing time
of CRIRES to reach the same SNR for
similar flux levels and spectral resolution.
The closest T dwarf to the Sun (3.26 pc),
e Ind B, is an ideal system to improve our
understanding of substellar objects by
detailed spectral measurements. The two
members of this extreme cool brown
dwarf binary have spectral type T1 and
T6.5 and masses of ~ 47 and 28 MJup,
respectively. The CRIRES spectroscopy
campaign by Seifahrt [ID 21] concentrat-
1.5
i = 20˚
R min = 0.10 AU
R max = 10.00 AU
1.0 α = – 2.0
0.5
0.0
– 0.5
– 50
Velocity (km s –1)
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1301 1302
ed on a K i line at 1252.5 nm, water va-
pour and methane (CH4 ) features at wave-
lengths close to the peak flux of T dwarfs.
For the first time in natural-seeing ob-
servations, spectra of both binary com-
ponents, separated by ~ 0.9?, have been
obtained with spectral resolution of
R ~ 50 000 (Figure 4). The S/N of the spec-
tra are between 20–60 (per pixel) which
is in close match to the predictions of the
CRIRES exposure time calculator (ETC).
Detailed comparison of the spectra to
synthetic atmospheric models as well as
to existing T dwarf models are in pro-
gress. In particular, the relative radial ve-
locity of e Indi Ba-Bb will improve their
orbital parameters.
Precise radial velocity (RV) measure-
ments are important in many astrophysi-
cal objects. In particular, they have led
to the discovery of more than 200 extra-
solar planets. Carrying out equivalent
measurements at infrared wavelengths
Figure 3: H2 n = 1−0 S(1) line detected in LkHa 264
at 2 121.8 nm and model (dashed) assuming that the
emission originates in a circumstellar disc. The fit
parameters R min and R max, the inner and outer radius
of the emitting region, a, the power law exponent
of the intensity (I(R) ∝ R a) and i, the inclination angle
(Carmona et al.), are listed.
0 50
Figure 4: Cut–out of the high-resolu-
ε Indi Ba (T1.0) tion spectrum of the binary compo-
nents e Ind Ba (top) and Bb (bottom).
Lines of water vapour (H2O) and
methane (CH4) are marked. Note the
increase in line strength for both spe-
H 2O cies between the T1 dwarf (Ba) and
CH 4 the T6.5 dwarf (Bb) and the strong line
broadening by the high projected
rotational velocity v sin i ~ 28 km/s
(Seifahrt et al.).
ε Indi Bb (T6.5)
1303 1304 1305 1306
Wavelength (nm)
has several advantages: indeed, late-
type stars and brown dwarfs are much
brighter at infrared wavelengths while
the RV-jitter caused by stellar activity is
about one order of magnitude smaller
in the infrared than in the optical regime
(Martins et al. 2006). Günther [ID 6] ob-
served the M0V star GJ9847 in two dif-
ferent nights, obtaining spectra with the
N2O-absorption cell, as well as tem-
plate spectra without. Simulations predict
that the simultaneous wavelength refer-
ence allows a radial velocity precision of
better than 30 m/s to be achieved. Even
though improvements in the data analysis
are still possible this accuracy is already
achieved. The basic idea of the procedure
is the same as for the I2-cell method in
the optical, which has been used for many
years for planet detection. The main dif-
ference is that the influence of the telluric
absorption lines has to be carefully re-
moved. This is achieved by dividing the
spectrum by a reference star free of
The Messenger 128 – June 2007 19
Telescopes and Instrumentation
absorption features, e.g. a B-type star.
This technique is illustrated in Figure 5 for
the M0V star GJ 9847.
Normalised flux (F λ + offset)
The SV programme by Moutou [ID 16]
aimed at detecting the absorption signa-
ture of a planetary atmosphere, during
a transit of the extrasolar planet in front of
its parent star. This requires high spec-
tral resolution to deblend the planetary
features from the stellar ones at a certain
epoch. For such transiting systems, one
expects to detect a K i absorption feature
in the near-IR. Unfortunately during the
execution of the programme, the CRIRES
observing sequence could not fully meet
the time critical window that lasted a little
more than two hours. The data however
show differences from one spectrum to
the other and a longer lasting sequence
is allocated in P79.
Massive stars, Asymptotic Giant Branch
stars and Galactic structure
The Pistol star, located close to the Gal-
actic Centre, is the most luminous star in
the Galaxy. Unless it is actually made up
of various components, it is also the best
candidate to be the most massive star.
The very high resolving power of CRIRES
was used by Martins [ID 14] to investigate
the presence of a spectroscopic com-
panion, as well as to constrain the abun-
dance of a few key elements (Si, Mg and
Fe). Data were obtained during the Febru-
ary run and are currently being analysed
with atmospheric models. They will re-
veal whether the Pistol star is truly a very
massive star, and will provide the ratio
of alpha elements to Fe, a quantity which
critically depends on the chemical en-
richment history and the IMF in the Gal-
actic Centre.
Information on star interiors can be ob-
tained through the determination of the
frequencies of pulsation modes as they
emerge at the stellar surface, a technique
known as astroseismology. Amado et
al. [ID 1] tested the feasibility of astroseis-
mology in the NIR. For 3.5 hours, they
continuously observed the strongly pho-
tometric d Scuti variable star, 1 Mon and
searched for the presence of line profile
or equivalent width variations due to the
pulsations.
20 The Messenger 128 – June 2007
Siebenmorgen R. et al., First Results from CRIRES Science Verification
Science target: GJ9847 (M0V)
2.0
Telluric transmission: (standard star 105 Aqr, B9V)
1.5
Gas cell: N 2O, first overtone, R(30)–R(0) and P(1)–P(2)
1.0
0.5
Final composite spectrum
0.0
2 254 2 256
O-type stars are among the brightest and
most luminous stellar components of a
galaxy. Through their radiation and wind
momentum, they largely influence their
surroundings as well as their host galax-
ies. However, our knowledge of these
objects is still fragmentary. One of the
currently most critical questions is related
to the exact properties of their stellar
winds, which determine their mass-loss
rates, and thus their evolution to become
supernovae. CRIRES allows to signifi-
cantly increase the number of accessible
diagnostic lines. Sana [ID 20] acquired a
high S/N ratio spectrum of z Pup, an O4
supergiant and one of the closest O-type
stars. The comparison of the observed
line profile with the one predicted from
atmospheric models should provide tight
constraints on its wind properties and,
more specifically, on its clumping factor.
Using ESO and IUE archives, a multi-
wavelength analysis covering the whole
spectral range from the UV to the near-
IR domain is foreseen. The complemen-
tarity of the diagnostic lines in these va-
rious wavelength ranges should help
to obtain a self-consistent view of the at-
mosphere of this massive star.
Nissen [ID 17] choose extreme metal-
poor stars in the Galactic halo to study
the abundance of Sulphur, an a-element,
considered an important diagnostic
for galactic chemical evolution (see the
article by Nissen et al. on page 38). In-
frared observations of atomic lines in the
near-infrared are an important comple-
ment to visible (UVES) spectra, as they
can be less sensitive to details of the at-
mospheric models and may have sub-
stantially larger oscillator strengths. An
impressive spectrum with a S/N of 330
Figure 5: From top to bottom and
shown as indicated: the normalised
spectrum of GJ9847; the telluric
standard star; the N2O gas cell; and
the final composite spectrum (Günther
et al.).
2 258 2 260 2 262 2 264
Wavelength (nm)
was achieved (Nissen et al. 2007 –
the first published refereed paper from
CRIRES).
The infrared absorption spectra of H+3
and CO were observed by Goto [ID 5] to-
ward the luminous infrared source in the
Quintuplet cluster GCS 3-2 and is shown
in Figure 6. Note the distinct absorption
profiles of H+3 R(1,1)l and CO R(1) along
the same line of sight. While CO mostly
probes the gas in molecular clouds in the
Galactic arms along the line of sight, the
H+3, responsible for R(1,1)l, occurs both in
arm clouds and in the gas at the Galac-
tic Centre region. The pedestal compo-
nent of R(1,1)l is almost indentical with the
broad absorption band of R(3,3)l obtained
at Subaru and Gemini South. R(3,3)l how-
ever exclusively samples the gas in the
Galactic Centre. The population of H+3 in
the metastable state (J,K) = (3,3), 361 K
above the ground state, attests to the
presence of warm (T ~ 250 K) and diffuse
gas (NH ~ 100/cm3) in the central mo-
lecular zone, which was unknown before
(Oka et al. 2005).
Uttenthaler [ID 27] obtained H- and K-
band spectra of a handful of bulge AGB
stars to determine the C/O and 12 C/ 13 C
ratios. These ratios are influenced by
a deep mixing event called the third
dredge-up (3DUP) which occurs on the
AGB. If 3DUP has occurred in a giv-
en star, the 12 C/ 13 C ratio should be high
(compared to the solar value of ~ 89),
otherwise it should be low (< 10). As an
example the CRIRES spectrum of M794
reveals a strong 12 CO 4–2 band head
but weak 13 CO lines (e.g. at 2 354.25 and
2 355.1 nm), so that this star has a rath-
er low 12 C/ 13 C ratio. Further support that

1.15
1.1
Relative intensity
1.05
1 pected from a non-processed old star,
0.95
0.9
GCS 3-2
– 200 –100 0 100
1.5
1.25
Relative intensity
1 the simultaneous observation of many
0.75 GCS 3-2
– 200 –100
VLSR (km/s)
M794 has not experienced a 3DUP is giv-
en by the non-detection of Technetium,
a radioactive indicator for this process.
Interestingly, as determined from UVES
spectra, this star shows Lithium in its
atmosphere, which is rather surprising for
a previously classified low-mass AGB
star. Most probably another mixing proc-
ess – called cool bottom processing – is
at work in this star, reducing the 12 C/ 13 C
ratio almost down to its equilibrium value.
Gustaffson [ID 7] studied the CNO abun-
dances in bulge giants, first, as a SV trial
run and subsequently in an allocated pro-
gramme in period P79. The team is a
large international collaboration including
members from Australia, Brazil, Chile,
France, Germany and Italy. Studying the
Milky Way Bulge is particularly impor-
tant and, as an ultimate goal, should an-
swer the question of whether it is the
Galaxy’s oldest region, perhaps together
with the Halo. Alternatively, it could have
been gradually formed, at least partially,
from stars and gas in the disc that mi-
grated inwards. These hypotheses can be
Figure 6: Infrared absorption spectra of H+3 R(1,1)l with an excess of O relative to Fe but
Subaru and CO R(1) towards the Quintuplet cluster GCS 3-2
compatible with the alpha elements (Mg,
observed with three different high-resolution spec-
trographs. The performance of IRCS at the Subaru Si and Ca). The C and N abundances
telescope (with spectral resolving power R = 20 000) are signs that matter exposed to the CN
and Phoenix at Gemini South (R = 75 000) is com- cycle (which conserves the sum of C
Gemini
pared to that of CRIRES at the VLT (R = 100 000).
and N nuclei) has been dredged up to the
The higher velocity resolution of CRIRES is the ob-
vious advantage among the other instruments on stellar surface. The sum of the abun-
VLT/CRIRES 8-m-class telescopes. dances of C, N, and O is close to that ex-
discriminated by obtaining the ages of formed with an excess of oxygen and
a considerable number of Bulge red giant alpha elements relative to iron. The pre-
R(1,1) I stars using the ratios of heavy element liminary elemental abundances are pre-
3.7155 µm abundances, in combination with their sented in Ryde et al. (2007), while a full
kinematics, as a chronometer. The main analysis will be presented in a forthcom-
reason for observing in the IR (see Ryde ing paper.
et al. 2005) is the much smaller interstel-
Subaru lar extinction so that the whole Bulge is
observable and not just a few windows Outside the Galaxy
transparent at optical wavelengths. Only
Gemini the near-IR offers all the indicators neces- Two extragalactic SV projects were exe-
sary for accurate determination of the im- cuted. Siebenmorgen [ID 24] tried to
portant C-N-O molecular equilibrium constrain variations in the fine structure
VLT/CRIRES in the atmospheres of cool stars, through constant, a = e2 /hc by observing the
[O iii] doublet redshifted to z ~ 2. The QSO
clean CO, CN and OH lines. The goal with 1148-001 was selected as it shows a
the SV run was to test the method. The bright doublet unresolved at low resolu-
CO R(1) result was successful, even though, only tion. Unfortunately, the lines are too
2.3433 µm three Bulge stars could be observed. broad when observed at a resolution of
Figure 7 shows the CRIRES spectrum of
0 100
one of them, Arp 4203. The wavelength
range 1531–1570 nm was recorded and Figure 7: Two sections of the CRIRES spectrum of
only a 1/4 of the spectrum is shown for Arp 4203 are presented (full line, black) together with
models with nitrogen increased by +0.4 dex (dashed,
clarity. From the derived abundances, blue) relative to the best fit, and when oxygen is in-
the findings are that, e.g., the giant star stead decreased by 0.2 dex (dotted, red). For clarity
Arp 4203 is depleted in C, enriched in N, the best-fit model is not shown (Gustaffson et al.).
1.0
Normalised flux
0.8
0.6
CN Fe Fe OH OH CN Fe+Si Ti CN Fe CN CN Ni Si+OH OH+CN CN
0.4
15 530 15 535 15 540 15 545 15 550 15 555 15 560 15 565
Wavelength (Å)
1.0
Normalised flux
0.8
0.6
CO band head
OH OH OH CN CN CN CN Fe Fe Fe OH+FeCN
0.4
15 565 15 570 15 575 15 580 15 585 15 590 15 595 15 600
Wavelength (Å)
The Messenger 128 – June 2007 21
Telescopes and Instrumentation
50 000 and no useful constraint could be
derived.
A second project focused on the ISM of
proto-galaxies at z > 2 (Smette [ID 25]).
The Na i and Ca ii absorption lines are
common features in the ISM of our Gal-
axy, but so far they have never been ob-
served in the ISM of objects with red-
shifts larger than one. CRIRES offers the
possibility to extend the study of the ISM
by means of these important lines for
damped Lyman-systems (DLA) at high z.
These lines provide important information
regarding the velocity structure of the
gas. Na i can be compared with the low-
ionisation lines seen in UVES spectra to
better constrain the ionisation conditions.
At low redshift, Ca ii is often seen in tur-
bulent material such as the one seen in
the merging galaxies. Comparison be-
tween Na i and Ca ii is also a very good
indicator of the clumpiness of the gas.
The QSO HE0251-5550, chosen from
the H/ESO DLA survey (Smette et al. in
prep.), has a hydrogen column density
log(NHI ) = 20.7 DLA in its spectrum and a
redshift of z = 2.3, allowing the search
for Na i and Ca ii doublets with CRIRES.
Although the object is bright enough in
R-band to close the AO loop, though with
little improvement in image quality, the
faintness of the QSO (H = 14.6) made
guiding somewhat difficult, as insufficient
flux is reflected back from the slit jaws to
the slit viewer detector. The spectra cov-
ering the Na i lines reached the expected
S/N but did not reveal the lines, probably
because of the low metal abundance of
this system. Unfortunately, the seeing
degraded during the promising, yet chal-
lenging, observations of the Ca ii lines and
the spectra have an unusably low S/N.
Despite the non-detection, these obser-
vations were a good test of the feasibility
to observe such ’faint’ targets with
CRIRES, in particular, when the high res-
olution allows one to avoid the effects of
telluric lines.
22 The Messenger 128 – June 2007
Siebenmorgen R. et al., First Results from CRIRES Science Verification
Calibration of CRIRES data: CRIRES and
the Earth’s atmosphere
Observing with CRIRES as well as cali-
brating and reducing CRIRES data re-
vealed multiple challenges that the sci-
ence verification programmes helped to
identify. A large number of the problems
that the CRIRES team met during the
execution of the programmes have been
solved in time for science operations,
while work is still going on for others.
For example, the high-resolution capabili-
ties of the instrument are a challenging
aspect when it came to the development
of a precise (s(v) ~ 70m/s), robust and
automatic wavelength calibration proce-
dure. The pipeline, now in use to assess
the quality control of the data, uses as
first guess the wavelength solution pro-
vided by a physical model of the instru-
ment, which is accurate to about 1–2 pix-
els. It then cross-correlates data with
other catalogues such as telluric features
computed by HITRAN or OH line lists.
In addition, CRIRES supports high-accu-
racy wavelength calibration by means
of gas cells and a ThAr hollow cathode
lamp (Kerber et al. 2007).
But an accurate and precise wavelength
calibration relies on a large number of
lines on each detector. Ideally, these lines
should be observed simultaneously with
the target. Such is the case for observa-
tions through gas cells. In several spec-
tral bands, the sky lines are very numer-
ous. However, little is known of their
apparent wavelength stability which could
limit their use for high-accuracy wave-
length calibration. Such a study – the fea-
sibility of using OH lines to calibrate
CRIRES data – is at the base of the pro-
posal of Lidman [ID 12], as this molecule
shows a rich ro-vibration spectrum.
Before CRIRES, OH sky line spectra had
only been systematically obtained with
spectral resolutions of 10 000 or less.
CRIRES spectra of the OH lines were tak-
en during evening twilight, which is the
time when they are at their brightest. In a
single 300-s exposure it is possible to
see up to a dozen lines on a single detec-
tor and up to four times that number
over the entire mosaic. In order to assess
their wavelength stability, data over sever-
al nights were obtained, immediately
followed by a ThAr spectrum. A quick in-
spection of the R ~ 100 000 spectra re-
veals that CRIRES resolves doublets
which are seen as single lines at spectral
resolution of 10 000 (see Figure 9 in
Siebenmorgen and Smette 2007). How-
ever, even at a resolution of 100 000
CRIRES does not resolve individual lines.
Also, a number of sky lines appear that
are not listed in available catalogues and
vice versa.
CRIRES is also helping ESO in prepara-
tion for the next generation of giant
telescopes. Kerber [ID 9] tested a method
to measure the atmospheric content of
precipitable water vapour (PWV) using the
equivalent widths of H2O absorption lines
in the near-IR which are imprinted onto
the spectra of early-type stars. An accu-
rate and efficient method to determine
the PWV is highly valuable for any ther-
mal IR instrument at an E-ELT. For quanti-
tative measurements, unblended water
lines are selected to have a minimum de-
pendence of their absorption coefficient
at a temperature of ~ 300 K. CRIRES
spectra of bright stars were observed on
several nights during SV. On some of
these nights routine measurements with
VISIR (see Smette et al. 2007) can be
used to establish the PWV from mid-IR
data allowing for comparison across
wavelength and method. Very acceptable
agreement between both methods and
instruments is found.
References
Carmona A. 2007, Ph.D. Thesis, University of
Heidelberg
Käufl H. U. et al. 2004, SPIE 5492, 1218
Käufl H. U. et al. 2007, The Messenger 126, 32
Kerber F. et al. 2007, ASP Conference Series 364,
ed. C. Sterken, 461
Martins F. et al. 2006, ApJ 644, L75
Nissen et al. 2007, A&A, accepted,
astro-ph/0702689
Oka et al. 2005, ApJ 632, 882
Roueff et al. 2006, A&A 447, 963
Ryde N. et al. 2007, astro-ph/0701916
Ryde N. 2005, in: “High Resolution Infrared Spec-
troscopy in Astronomy”, ESO Astrophysics
Symposia, (eds.) H. U. Käufl, R. Siebenmorgen
and A. F. M. Moorwood, Springer, 365
Siebenmorgen R. and Smette A. 2007, CRIRES
User’s Manual, http://www.eso.org/instruments/
crires/doc/
Smette A., Horst H. and Navarrete J. 2007,
Proceedings of the 2007 ESO Instrument Cali-
bration Workshop, Springer-Verlag, in press,
http://www.eso.org/gen-fac/meetings/cal07/
presentations/smetteCal07.pdf
Telescopes and Instrumentation

The First Active Segmented Mirror at ESO

Frédéric Gonté, Christophe Dupuy, The design and integration phases of onal mirrors have to be aligned with a
Christoph Frank, Constanza Araujo, APE, which started in 2005, have been precision better than 15 nm rms. It was
Roland Brast, Robert Frahm, completed and the test phase will start clearly not feasible to produce a scaled-
Robert Karban, Luigi Andolfato, in June 2007 during which time APE down version of the full primary mirror.
Regina Esteves, Matty Nylund, will be installed at the focus of one of the However, from a statistical point of view,
Babak Sedghi, Gerhard Fischer, Lothar VLT unit telescopes in 2008. a mirror with approximately 50 segments
Noethe, Frédéric Derie (all ESO) is already representative of a mirror with
Initially the APE team wanted to contract many more segments in terms of the
the design and manufacture of the ASM study of issues like alignment algorithms
The Active Phasing Experiment (APE) is to a private company. However, when or the effect of misalignments on image
part of the Extremely Large Telescope no company could be found which could quality. The main requirements for the
Design Study which is supported by the meet the rather stringent requirements, ASM can be summarised as:
European Framework Programme 6. it was decided to develop the ASM in- – 61 segments in four rings around the
This experiment, which is conducted in house, involving ESO groups in Integra- central segment
collaboration with several partners is tion, Optics, Electronics, Software and – Segment size 17 mm to minimise the
a demonstrator to test and qualify newly- the ELT Project Office. size of the re-imaged pupil and
developed phasing sensors for the align- the relay optics on the optical bench
ment of segmented mirrors and test – Three degrees of freedom for rigid body
the phasing software within a telescope Design and integration movements, that is piston, tip and tilt
control system to be developed for a for each segment
future European Extremely Large Tele- The current design of the primary mirror – Precision of displacement better than 2
scope. The segmentation of a primary of the European Extremely Large Tele- nm (similar to an ELT)
mirror is simulated by a scaled-down scope consists of 984 hexagonal seg- – Range of displacement more than
Active Segmented Mirror of 61 seg- ments (Gilmozzi and Spyromilio 2007). 15 μm (similar to an E-ELT)
ments which has been developed in- Each segment has a diameter of around – Size of the gaps between segments
house. 1.5 m and the size of the gaps between between 80 and 150 μm (scaled-down
the segments is approximately 4 mm. from the gap size in the ELT primary)
In order to achieve the performance re- – Surface form quality better than 15 nm
Background quired for high-resolution imaging with RMS (similar to the required surface
extreme Adaptive Optics (AO), the hexag- quality of a real segment with active
In order to gain experience with the phas- shape control)
ing of segmented mirrors, in particular – Operational temperature between 0˚ C
Photo: H. H. Heyer, ESO

to develop new optical phasing sensors and 25˚ C

and to test the interface with the rest of – Possibility to exchange a segment in
the wavefront control system, ESO de- case of failure.
cided to perform an optical-bench type – Optical fibres to be inserted in two of
experiment, named the Active Phasing the hexagonal mirrors (central mir-
Experiment (APE). This project is part of ror and one of the perimeter mirrors) to
the Extremely Large Telescope Design align APE with high efficiency
Study (ELTDS) which is supported by the
European Framework Programme 6. The A segment unit, shown in Figure 1, con-
project is performed in close collabora- sists of a base, three actuators, three
tion with the Istituto Nazionale de Astro- springs and a hexagonal mirror. The 61
fisica di Arcetri, the Instituto de Astrofísica modules were assembled on a base plate
de Canarias (IAC) and the Laboratoire (Figure 2). Since the precision of position-
d’Astrophysique de Marseille (LAM), and ing the surfaces of the hexagonal mir-
the industrial firm Fogale. The core optical rors relative to the base plate is better
element of APE, simulating the segment- than 15 µm, it is necessary to be able to
ed primary mirror in a large telescope, is actively align all the mirrors. Among the
a scaled-down Active Segmented Mirror set of candidates for the commercially
(ASM) with 61 segments which can all available actuators, only piezo actuators
be controlled in piston, tip and tilt. The could fulfil the ASM requirements. Since
pupil of the telescope is re-imaged onto this type of active segmented mirror is
the ASM. The optical beam is then re- the first of its kind it was decided to start
adapted to behave like the VLT output development in mid-2005 with a proto-
beam and distributed to the four phasing type of only seven segments.
sensors to be tested in the experiment.
Figure 1: Each segment is integrated separately. It is
composed of one base, three piezo-actuators, three
springs and a hexagonal mirror.

The Messenger 128 – June 2007 23

Telescopes and Instrumentation Gonté F. et al., The First Active Segmented Mirror at ESO

Figure 3: Integration of

Photo: H. H. Heyer, ESO

the ASM (having a plug
and play concept) by
Christophe Dupuy. Each
segment is added one
by one on a support
polished plate which has
a surface ­flatness better
than 2 microns.

Figure 2: Mechanical design of the ASM made by

Christoph Frank. The 61 segments are seen with
the three actuators per segment, the module base,
the support plate and the connectors behind
the support plate. The fibres inserted in the central
mirror and one of the mirrors on the right side of the Figure 4: Photo of the
Photo: H. H. Heyer, ESO

ASM can also be seen. fully assembled ASM.

The mechanical design has been made

by Christoph Frank. The integration has
been conducted by Christophe Dupuy for
the opto-mechanical parts (see Figure 3),
by Roland Brast for the drive electron-
ics and by Robert Frahm for the software.
The tests were performed by Constanza
Araujo. Information obtained during
the tests of the seven-segment prototype
was used to improve the final design.

obtained by an Internal Metrology sys-

Current tests tem which consists of a two-wavelength
interferometer developed by Fogale.
Several features have been tested on the The seven-segment prototype has been
fully integrated ASM (Figure 4), for exam- used to test the closed-loop system
ple fatigue of the components, eigen- and achieved a performance better than
frequencies, speed and hysteresis of the 5 nm rms for the alignment errors at the
positioning, drift of an actuated position, intersegment borders.
resolution and accuracy of the position-
ing, quality and cosmetics of the surfaces The fully integrated ASM has now fulfilled
of the segments. The fully integrated ASM and, in some areas, even surpassed the
has fulfilled and, in some cases, even sur- specifications. The success of this realisa-
passed the specifications. Figure 5 shows tion now provides the possibility to fully
an interferogram of the surface of the test future phasing sensors of the E-ELT
ASM while being tested. The tips and tilts within the Active Phasing Experiment.
of the segments have been selected such
that the letters of the acronym ASM ap-
pear in the interferogram. References Figure 5: Interferogram of the surface of the ASM.
Some segments have been tip-tilted to create fringes
Gilmozzi R. and Spyromilio J. 2007, The Messenger on some selected segments to write on it the
During the APE experiment, the ASM will 127, 11 acronym ASM.
be driven in closed loop based on signals

24 The Messenger 128 – June 2007

Telescopes and Instrumentation
Progress of the ALMA Project
Christoph Haupt, Hans Rykaczewski
(ESO)
An overview of the current status of the
ALMA project and the progress over the
last two years is presented. The main
focus is on the European deliverables,
but the contributions of North America
and Japan are also mentioned.
Science Objectives of ALMA
The Atacama Large Millimeter and Sub-
millimeter Array (ALMA) will provide an
unprecedented combination of sensitivity,
angular resolution, spectral resolution,
and imaging fidelity at the shortest radio
wavelengths for which the Earth’s atmos-
phere is transparent. The three major
scientific goals for ALMA can be summa-
rised as: to detect spectral line emission
from CO or C ii at a redshift of z = 3 in
less than 24 h of observation; to image
gas kinematics in protostars and proto-
planetary discs at a distance of 150 pc,
enabling the study of the physical, chemi-
cal and magnetic field structures; to pro-
vide images with an angular resolution of
0.1?.
Developments over the last two years
The agreement to construct ALMA was
signed in 2002 by the North American
community, represented through the Na-
tional Science Foundation (NSF), and
the European community, represented
through ESO. After three years of detailed
design studies and prototype research,
it became evident that the project could
not be realised within the original time-
scale and within the available funds fore-
seen at the time of signing the agree-
ment. The management and scientists
involved in ALMA were therefore charged
to develop a new configuration of the
project within their foreseen budget, while
still maintaining the prime scientific objec-
tives.
During this period in-depth studies and
large efforts were made to define the “re-
baselined” ALMA project. A major ele-
ment in cutting the cost was the reduc-
tion of the number of antennas from 64 to
50 (with an ultimate goal of 64). Fortuna-
AOS Site (43 km)
tely, the results of testing of the proto-
types for the receiver cartridge showed
that several of the receivers significantly
exceeded their noise temperature re-
quirements, thus partially offsetting the
loss in sensitivity through the reduced
number of antennas. Independent advi-
sory and review committees, focusing on
the technical performance as well as on
the cost of the project, confirmed the sci-
entific importance of the modified ALMA
project, based on the new affordable
costs. By the end of 2005 the re-base-
lined project and the new budget were
approved by the ESO Council. The North
American partners received approval
from their Funding Agency (NSF) by the
middle of 2006.
In parallel, Japan, through the National
Astronomy Observatory of Japan (NAOJ),
continued to define and formulate their
participation in the ALMA project. The
European and North American partners
in ALMA spent a considerable amount of
time with their Japanese partners in iden-
tifying the Japanese participation and
reviewed various subsystems, in particu-
lar correlator, receivers and antennas.
Details on the partnerships were defined
and an official, trilateral agreement be-
tween ESO, the NSF, and the National In-
stitute for Natural Sciences (NINS, Japan)
was signed in summer 2006. NAOJ will
provide four antennas of 12 metres diam-
eter, twelve antennas of 7 metres diame-
ter, two receiver bands for all 66 anten-
nas of ALMA and the Atacama Compact
Array (ACA) Correlator. Approval of funds
required for the Japanese participation
Figure 1: The Road
between Chilean High-
way No. 23 and ALMA.
Photo: H. H. Heyer, ESO
OSF Site (15 km)
in ALMA is expected by summer 2007.
With the inclusion of the Japanese
partners ALMA becomes a truly global
astronomy facility.
The ALMA Site
The ALMA Array Operations Site (AOS)
will be located at a unique place: the Alti-
plano de Chajnantor, a plateau at an alti-
tude of 5 000 metres in the Atacama Des-
ert in Chile. This location was selected for
reasons of dryness and altitude.
The ALMA Operations Support Facili-
ties (OSF), located at an altitude of about
2 900 metres and 28 km away from the
AOS will be the base camp for the every-
day, routine operation of the observatory.
Constructing and operating the ALMA
Observatory within the ambitious sci-
entific goals and the unprecedented tech-
nical requirements in this environment,
with its harsh living conditions, will be one
of the real challenges.
Construction of the OSF and AOS sites
and their access required substantial
effort from the ALMA project. The OSF
site, located at 2 900 metres altitude, is
about 15 km away from the nearest pub-
lic road, the Chilean highway No. 23
and the AOS is another 28 km away from
the OSF site. A road of 43 km length
was constructed to access the OSF and
AOS (Figure 1), with sufficient width to
regularly transport the antennas between
the OSF and AOS for maintenance.
The Messenger 128 – June 2007 25
Telescopes and Instrumentation Haupt C. and Rykaczewski H., Progress of the ALMA Project

Figure 2: Construction
work on the OSF
Technical Buildings.

The OSF is the centre of activities of the

ALMA construction project. Presently
all ALMA Site contractors and their staff
are accommodated at the OSF. Work is
organised in 20 day working/10 day rest
periods. Special camps have been
erected and can now accommodate the
maximum required capacity of 500 work-
ers.

During the construction period, the OSF

will become the focal point of all antenna
Assembly-Integration-Verification (AIV)
activities. Antenna assembly will be done
at the OSF site at three separate areas,
one each for the antennas provided by
North America (Vertex), Japan (Melco)
and Europe (AEM Consortium). AIV activ-
ities will be carried out at the OSF, after
preliminary acceptance of the antennas,
and prior to moving them to the AOS.

Ultimately, the OSF and its Technical Fa-

cilities will become the centre of all day-
to-day scientific activities. During the op-
erations phase of the observatory it will
be the workplace of the on-site astrono-
mers and of the technical teams respon-
sible for maintaining proper functioning of
the ALMA telescope and its equipment.

ESO signed the contract for the construc-
tion of the OSF Technical Facilities in
August 2006. Construction work has ad-
vanced (see Figure 2) and during this
time no major problems or delays have
occured. As specified in the construc-
tion schedule, foundations have been
prepared and the superstructures of the
buildings were completed in March 2007.
Provisional Acceptance of all facilities
is foreseen for the first quarter of 2008.
The AOS Technical Building (Figure 3) will
be delivered by the North American ALMA
partner. Inside the building, construction
work and furnishing is expected to be
completed by summer 2007. Human op-
erations at the AOS will be limited to
an absolute minimum, due to the high alti-
tude. The AOS Technical Building will
house the Back-End electronics and the
Correlator. Digitised signals received
from the radio telescopes will be proc-
essed here and further transmitted to the
data storage facilities located at the
OSF. The AOS building will become us-
able for installing technical equipment in
June 2007.
Figure 3: The AOS
­Technical Building.
The ALMA Antennas
Major performance requirements of each
antenna are 2? absolute pointing over the
whole sky, 0.6? offset pointing, a 25 mi-
cron rms overall surface accuracy and
the ability to fast switch over a 2-degree
range in less than 1.5 seconds. These re-
quirements are at least comparable to
those of existing submillimetre radio tele-
scopes; however all of these, except the
APEX antenna which is very similar to an
ALMA prototype antenna, are protected
from the weather by shelters.
The antennas are critical for the ALMA
project and their quality and performance
are a major contributor to the overall
functionality of ALMA. In view of the tech-
nical criticality of the antenna specifica-
tion, three prototypes were supplied by
the AEC Consortium (procured by ESO),
Vertex RSI (procured by NRAO for North
America) and Mitsubishi Electrical Com-
pany (procured by NAOJ, Japan). All three
prototypes were extensively tested at the
ALMA Test Facility in Socorro, New Mex-
ico (see the image on the front cover).
Various groups of international experts,
both internal and external to ALMA, re-
viewed the performance of the prototype
antennas and concluded that the per-
formance expected at the ALMA site con-
forms to the technical requirements for all
three designs.
Following a Call for Tender for the anten-
nas, the North American partners of the
ALMA project, through Associated Uni-
versities Incorporated (AUI), signed a con-
tract to supply up to 25 antennas, with
options to increase the contract to 32 an-
tennas, with Vertex RSI on 11 July 2005.
Similarly, on 6 December 2005, the ESO
Director General signed a contract (see
ESO PR 31/05) with the AEM Consor-
tium (Alcatel Alenia Space France, Alcatel
Alenia Space Italy, European Industrial
Engineering S.r.L., MT Aerospace) for the
supply of 25 ALMA antennas (as the Eu-

26 The Messenger 128 – June 2007


ropean share of the project), also with
options to increase the number of anten-
nas to 32. The four total power antennas
of 12 m diameter, equipped with nuta-
tors, to be provided by Japan, have been
ordered from Mitsubishi Electrical Com-
pany. The twelve remaining antennas of
7 m diameter will be ordered in the course
of the year 2007 by NAOJ.
The first antenna to be supplied by Vertex
RSI is expected to be ready for provision-
al acceptance in Chile in the second half
of 2007. The first antenna to be supplied
by the AEM Consortium is expected by
the third quarter of 2008. Despite the de-
layed delivery of the AEM antennas, both
suppliers are expected to deliver their 25th
antenna by the end of 2011. Acceptance
for the first antenna from Japan is ex-
pected to take place in December 2007.
The Preliminary Production Design Re-
view for antennas to be produced by
Vertex RSI was held in September 2006,
the corresponding review for AEM an-
tennas was held at the end of January
2007.
The ALMA Antenna Transporters
The antenna array at an altitude of
5 000 m can be reconfigured by relocat-
ing antennas on the stations located on
the Chajnantor plateau. The configura-
tions can be changed from a compact
one, in which all antennas operate within
an area of 160 m × 250 m, to an extended
confi-guration for which the maximum
separation between antennas reaches
about 15 km. In order to move antennas,
each with a mass of around 100 tons, the
ALMA project has designed a special
transport vehicle (Wilson 2006). Two vehi-
cles will be needed to cover the opera-
tional relocation requirements. These
transporters will move antennas between
the assembly and maintenance area at
the OSF and the AOS and at the AOS be-
tween different antenna stations for re-
configuration of the array. By means of
the transporter (Figure 4) the antennas
can be positioned to an accuracy of a
few millimetres on the high-precision sta-
tion interface.
The mass of the antennas, their high-pre-
cision and the hostile, high altitude envi-
Figure 4: ALMA Antenna
Transporter.
ronment impose severe requirements
on these vehicles. Each transporter will
have a mass of about 150 tons, and
dimensions of about 10 × 15 × 6 m (width
× length × height). The contract for the
production of these truly unique tranport-
ers was signed by ESO with Scheuerle
Fahrzeugfabrik GmbH (Germany) on
22 December 2005. The first of the trans-
porters will roll out of the hangar and will
be tested in summer 2007 (see Figure 4)
to be transported and delivered to the
OSF in the fourth quarter of 2007. The
second vehicle is expected to be deliv-
ered about six months after the first one.
The ALMA Front-End
The ALMA Front-End system is the first
element in a complex chain of signal re-
ception, conversion, processing and re-
cording. The Front-End is designed to re-
ceive signals in ten different frequency
bands (Table 1). In the initial phase of op-
eration the antennas will be equipped
with six bands. These are Bands 3, 4, 6,
7, 8 and 9. It is planned to equip the an-
tennas with the missing bands at a later
stage of ALMA operation.
The ALMA Front-Ends are superior to al-
most all existing systems. Indeed, devel-
opment work for the ALMA prototypes
has also led to improved receivers for ex-
isting millimetre and sub-millimetre obser-
vatories.
Photo: Scheuerle Fahrzeugfabrik GmbH
They are comprised of numerous elements
(Figure 5), produced at different locations
in Europe, North America and Japan.
In the initial phase of construction ALMA
decided, after the prototyping and de-
veloping stage, to build a set of eight pre-
production units before moving to full pro-
duction. This initial phase started in the
years 2004 and 2005. Some components
were successfully and completely pre-
produced by the end of 2006 and the
rest are expected to be completed in the
course of 2007.
The largest single element of the Front-
End system is the cryostat, a vacuum
vessel with the cryo-cooler attached. The
cryostats will house the receivers, which
are assembled in cartridges and can re-
latively easily be installed or replaced. The
corresponding warm optics, windows
and IR filters were delivered by the Institut
de Radio Astronomie Millimétrique (IRAM,
France). The operating temperature of the
cryostats will be as low as 4 K.
ESO and the Rutherford Appleton Labo-
ratory (RAL, UK) launched a development
and pre-production programme for the
manufacture of eight completely opera-
tional cryostats. Six cryostats (Figure 6)
have been fully assembled. The remaining
two units are in manufacture and the Cry-
ostats production contract was signed in
April 2007.
A development and pre-production
phase was also launched for the six dif-
The Messenger 128 – June 2007 27
Telescopes and Instrumentation Haupt C. and Rykaczewski H., Progress of the ALMA Project

Table 1: The ten frequency bands of ALMA.

ALMA Band Frequency Range (GHz) Receiver Noise Temperature (K) To be produced by 1 Mixing scheme
over 80 % of the RF Band at any RF Frequency
1 31.3–45.0 17 26 to be decided Single Side Band
2 67–90 30 47 to be decided Single Side Band
3 84–116 37 60 HIA Dual Side Bands
4 125–163 51 82 NAOJ Dual Side Bands
5 163–211 65 105 OSO (6 only) Dual Side Bands
6 211–275 83 136 NRAO Dual Side Bands
7 275–373 147 219 IRAM Dual Side Bands
8 385–500 196 292 NAOJ Dual Side Bands
9 602–720 175 261 NOVA Double Side Band
10 787–950 230 344 NAOJ (to be confirmed) Double Side Band

ferent receiver cartridges (Table 1). The 1

IRAM Institut de Radio Astronomie Millimétrique NOVA Nederlandse Onderzoekschool voor
(Grenoble, France) Astronomie (Groningen, the Netherlands)
technical specification of the various
HIA Herzberg Institute of Astrophysics (Victoria, NRAO National Radio Astronomy Observatory
receiver cartridges are demanding, set Canada) (Charlottesville, USA)
at the state of the art or even beyond NAOJ National Astronomical Observatory of OSO Onsala Space Observatory/Chalmers
at the time of definition. The development Japan (Mitaka, Japan) Univeristy (Onsala, Sweden)
programmes were successful, meeting
essentially all of the requirements – and
sometimes substantially exceeding
them – specifically for the receiver noise
temperatures. This is a very important
achievement as to some extent the better FESS Tertiary Mirror Optics Tertiary Lens Optics
Figure 5: Schematic
of the ALMA Front-End
noise performance and higher sensitivity (Bands 3 and 4) (Bands 1 and 2)
System.
partially compensate the loss in collecting
area.

In the frame of the European FP6 pro- WVR

gramme, ESO is leading a group of Euro-

pean institutes to develop and build six Cryocooler Vacuum Vessel
Band 5 receiver cartridges and to de-
velop associated software. This project
was approved by the European Commu- IF Switch Module Bulkhead
IF Switch Control Module AMB Bus × 2
nity at the end of 2005. In parallel, NAOJ FEMC Module IF Outputs (4)
Cryostat M&C Module DC Power in (#2)
is leading activities related to the devel- DC Power in (#1)

opment of Band 10 receivers. The Band 9 DC Power Districution Modules (10) Photonic Reference (Source)
Photonic Reference (Return)
cartridge was built by NOVA in the Neth- DC Supply M&C Module
LO Offset LO Offset Reference Input
erlands and is shown in Figure 7. The re- Distribution Module

ceivers for Band 10 are still in an R&D

phase; it is expected that in 2008 a final
decision will be made on implement-
ing this ALMA band with the highest fre-
quency. Figure 6: ALMA
Cryostat pre-production
at RAL/UK.
Pre-production of the cryogenic low-noise
amplifiers for Band 7, 4–8 GHz, and for
Band 9, 4–12 GHz, receivers has been
successfully completed by Centro Astro-
nómico de Yebes (CAY, Spain). Ampli-
fier series production for Band 7 com-
menced in February 2007 by Spanish
industry.

Water Vapour Radiometers (WVRs) are

needed to provide a correction of the
phase fluctuations due to atmospheric
water vapour fluctuations. The develop-
ment of two prototype WVRs, at Cam-

28 The Messenger 128 – June 2007


Figure 7: ALMA Band 9
Cartridges ready for
delivery, at NOVA in the
Netherlands.
bridge University (UK) and Onsala Space
Observatory (Sweden), was completed
and both underwent intensive testing at
the Sub-Millimeter Array (SMA) on Mauna
Kea (Hawaii). The performance of both
prototypes meets the requirements and
demonstrated the atmospheric correction
method. The Call for Tender for the final
detailed design and full WVR production
was issued in February 2007 to indus-
try in ESO member states and production
should commence in mid-2007.
The ALMA Back-End
The ALMA Back-End systems deliver sig-
nals generated by Front-End units in-
stalled in each antenna to the Correlator
installed in the AOS Technical Building.
Signal processing and data transfer is
schematically shown in Figure 8. In each
of two orthogonal linear polarisations
an intermediate frequency (IF) of 8 GHz
bandwidth can be processed and re-
corded at any time. Analogue data, pro-
duced by the Front-End electronics, are
processed and digitised before being for-
matted for the transmission through the
optical transmitter units and multiplexers.
All these elements are installed in the
receiver cabins of each antenna. Optical
signals are then sent through optical fi-
bres to the AOS Technical Building. The
total distance in the most extended an-
tenna configuration is about 15 km. At the
AOS Technical Building the incoming op-
tical signals are de-multiplexed and de-
formatted before entering the Correlator.
Figure 8: Schematic of ALMA Signal
Processing and data transfer from
the Front-End to the Correlator. Parts
to be provided by ESO are shaded in
dark blue.
Antenna
Front-End
IF Processing
8 × 2–4 GHz sub-bands
Digitizer
Local Oscillator
8 × 4 GHz – 3 bit ADC
8 × 250 MHz, 48 bit cut
Data Encoder
12 × 10 GHz
12 Optical Transmitters
12 1 WDM Optical MUX
The European deliverables in the ALMA
Back-End project are a number of com-
ponents, which are produced by several
European institutes, working closely with
ESO, and NRAO. These deliverables are:
digitizer chips and digitizer assembly; di-
gitizer clock assembly; optical data trans-
mission system; fibre patch panel; optical
multiplexers (MUX) and de-multiplexers
(De-MUX); Erbium-Doped Fibre Amplifiers
(EDFA), and photonic local oscillator pho-
tomixers.
Development and pre-production of
these components has been successfully
completed; the components will be inte-
grated at NRAO in Socorro and installed
in the European and North American pro-
totype antennas for tests at the ALMA
Test Facility ATF.
An example of one of the various Back-
End components produced in Europe is
the 4 GHz digitizer in its final layout (see
Baudry et al. 2006, Figure 3). The clock
rate is 4 GHz, allowing an input bandwidth
of 2 GHz. The digitization uses three lev-
els to preserve the signal-to-noise ratio.
During the prototype and development
phase, the initial layout was optimised in
order to reduce the number of parts and
the assembly costs. The final digitizers
show an improved performance and
a higher reliability. This work was carried
out in close collaboration between ESO
and the University of Bordeaux. The first
units have been shipped to Socorro
for final acceptance and integration with
tests at the ATF.
AOS Technical building
Photo-
Correlator
mixers
Tunable Filter
Digital De-Formatter
Digitizer
Clock
Optical De-MUX
and Amplifier
Fibre Patch Panel
From 175 stations
to 64 DTS Inputs
Optical Fibre
The ALMA Correlator
The ALMA Correlator, to be installed in
the AOS Technical Building, is the last
component in the receiving end of the
data transmission. It takes as input the
digitized signals from the individual an-
tennas and outputs amplitude and phase
on all of the interferometer baselines in
each of a large number of spectral chan-
nels. It is a very large data-processing
system, composed of four quadrants,
each of which can process data coming
from up to 16 different antennas. The
complete correlator will have 2 912 print-
ed circuit boards, 5 200 interface cables,
and more than 20 million solder joints.
The first quadrant was completed at
NRAO in the third quarter of 2006. Work
on the second quadrant is progressing
on schedule.
Integral parts of the Correlator are Tuna-
ble Filter Bank (TFB) cards, which allow a
major increase in the flexibility by sub-
dividing the frequency range into 32 inde-
pendently configurable sub-channels.
Four TFB cards are needed for the data
coming from a single antenna. The TFB
cards have been developed and opti-
mised by the University of Bordeaux over
the last few years. Prototypes and pre-
production units were extensively tested
and their performance was critically
reviewed in the first half of 2006. In the
meantime, series production is pro-
gressing and the first batch of 108 out of
560 TFB cards has been produced.
The Messenger 128 – June 2007 29
Telescopes and Instrumentation
System Engineering and Integration
An important ALMA-wide activity is the
System Engineering and Integration
(SE&I) task, which covers coordination of
activities across the project. One of the
prime objectives of SE&I is to ensure that
modules and equipment produced for an
ALMA subsystem, often in different loca-
tions of the world, fit physically and func-
tionally together. SE&I is also responsible
for the ALMA system design, technical
budgets and for ensuring the system inte-
grity, and hence has to ascertain that all
interfaces between the various subsys-
tems are completely understood, well de-
fined, rigorously reviewed and properly
documented.
In order to achieve the tasks mentioned
above, the SE&I team takes a central role
in the preparation and management of
project performance and engineering re-
quirements and Interface Control Docu-
ments ICD. SE&I is also very active in
organising, holding and chairing technical
reviews of the project, which could be
internal to ALMA or external, reviewing
the status and progress of the design
and manufacturing of ALMA components
by industry or scientific institutes. Product
and Quality Assurance is another task
and will receive a high visibility specifically
during the manufacturing phase.
The SE&I group is also leading the activ-
ities related to Prototype System Integra-
tion and the testing at the ALMA Test Fa-
cility (ATF) in Socorro (see front cover).
During the years 2007 and 2008 the ATF
will serve as a facility where ALMA hard-
ware and software will be integrated into
an operational system using the two pro-
totype antennas from Europe and North
America. In a joint effort with the ALMA
Science team, first fringes were recorded
in March 2007 (see ESO PR 10/07). The
ATF is also used for training integration
and operations personnel for Chile and
for qualifying equipment needed for ac-
ceptance tasks in Chile, such as the Opti-
cal Pointing Telescope and the Hologra-
phy System.
Another major activity for the coming
years is the Assembly, Integration and
Verification (AIV) at the OSF. Equip-
ment acceptance has already started and
AIV at the OSF will begin with the antenna
30 The Messenger 128 – June 2007
Haupt C. and Rykaczewski H., Progress of the ALMA Project
delivery in the middle of 2007 for which
preparations are currently ongoing.
Computing
The development of a unified software
system for the ALMA Observatory is an-
other important activity which requires
intensive interactions with all of the ALMA
engineering project teams. The nature
of software development calls for a com-
mon approach and policy from the very
beginning – not only with respect to the
engineering teams, but also between the
European and North American partners.
In contrast to some cases of hardware
oriented activities, where specialised in-
struments are provided, this project re-
quires entire software packages to be de-
veloped under a single management and
organisation.
ALMA Software Development is man-
aged as a truly integrated, trilateral orga-
nisation. The European, North American
and Japanese Executives have identified
individual work packages under a com-
mon leadership.
Work in Europe is organised through
ESO. The centre of European software
development and management is at
ESO’s headquarters in Garching. Many
software packages are developed by
ESO in collaboration with European in-
stitutes who have experienced and quali-
fied staff developing and testing the
required software. Table 2 gives an over-
view of software projects carried out by
various European institutes in association
with ESO.
Major computing subsystems for which
ESO has direct responsibility include:
– ALMA Common Software (ACS), the
infrastructure used by all other soft-
ware;
– Archive, both Front-End and science
archive, collecting not only science
data, but also any other information
used by the ALMA observatory soft-
ware;
– Integration and test activity, to get a
coherent and tested software system
out of the many subsystems;
– High Level Architecture activity, respon-
sible for the design of the ALMA soft-
ware and consistency of its interfaces;
Table 2: Software Development in European
Institutes.
Software Activity Institute 2
Science Software Requirements IRAM
Pipelining UKATC, MPIfR
Archiving Man. Un./JBO
Observing Preparation UKATC
Offline Data Processing MPIfR, CNRS
Telescope Calibration IRAM, IEM-CSIC
ALMA Common Software OAT
2
IRAM Institut de Radio Astronomie Mil-
limétrique (Grenoble, France)
UKATC United Kingdom Astronomy Technology
Centre (Edinburgh, United Kingdom)
MPIfR Max-Planck-Institut für Radioastronomie
(Bonn, Germany)
Man. Un. University of Manchester (Manchester,
United Kingdom)
JBO Jodrell Bank Observatory (Macclesfield,
United Kingdom)
CNRS Observatoire de Paris & Centre National
de la Recherche Scientifique (Paris,
France)
IEM-CSIC Instituto de Estructura de la Materia
(Madrid, Spain)
OAT Osservatorio Astronomico di Trieste
– E xecutive subsystem, which provides
the operator user interfaces;
– Observatory Support Software, which
provides the support tools needed by
time allocation teams and the support
for data packages;
– Software engineering activity, which
maintains the standards to be used in
hardware and software.
Figure 9 shows as an example the Opera-
tor Master Console. This is part of the
Executive subsystem, for which ESO is
responsible.
Acknowledgements
We would like to thank Claus Dierksmeier, Stefano
Stanghellini, Gie Han Tan, Fabio Biancat Marchet
and Gianni Raffi for helping to prepare this article.
References
Baudry A. et al. 2006, The Messenger 125, 37
Wilson T. 2006, The Messenger 123, 19
Haupt C. et al. 2006, Proc. SPIE 6271, 14

ALMA European Project Scientist Appointed
Tom Wilson (ESO)
The new ALMA European Project Scien-
tist is Dr. Leonardo Testi. He took up the
appointment in May 2007. Leonardo Testi
received his Ph.D. from the University
of Florence in 1997. Subsequently he was
a postdoctoral fellow at the Owens Valley
Radio Observatory of Caltech. In 1998
he joined staff of the Arcetri Astrophysi-
cal Observatory, and later on of INAF, for
which he also served on the Science
Council. Leonardo has been chair of the
European ALMA Science Advisory com-
mittee and a member of the ALMA
Science Advisory committee, so he well
knows the details of the project as well as
the science that can be carried out with
ALMA.
Figure 9: The
ALMA Operator
Master Console.
Leonardo’s main scientific interest is the
study of circumstellar discs around newly Photo: H. H. Heyer, ESO
formed stars. It is believed that many of
these discs of dust and gas will develop
into planetary systems, so such studies
are complementary to the optical searches
for extrasolar planets being carried out,
for example, with HARPS and the ESO
3.6-m telescope. The major drawbacks to
present-day studies of circumstellar discs
are sensitivity and angular resolution.
With ALMA, the sensitivity will be more
than 40 times better, and the angular res-
olution more than 5 times better, than
current instruments. Thus ALMA will allow
breakthroughs in this area, as well as for
studies of star formation, galaxy evolution
and dynamics, Solar System science and
cosmology. Dr. Leonardo Testi
The Messenger 128 – June 2007 31
Astronomical Science

FORS1 colour composite image of

the H ii region NGC 2081 in the
Large Magellanic Cloud taken with
the new blue CCD. Three filter im-
ages isolating emission lines of [O ii],
[O iii] and H-alpha were combined.
Astronomical Science

Hunting for Frozen Super-Earths via Microlensing

Jean-Philippe Beaulieu 1, 3 8
Boyden Observatory, University of of cold telluric planets. Its detection
Michael Albrow 1, 4 the Free State, Department of Physics, confirms the power of this method and,
Dave Bennett 1, 5 Bloemfontein, South Africa given our detection efficiency, suggests
Stéphane Brillant 1, 6 9
Astronomisches Rechen-Institut (ARI), that these recently-detected planets
John A. R. Caldwell 1, 7 Zentrum für Astronomie, Universität may be quite common around M stars,
Johannes J. Calitz 1, 8 Heidelberg, Germany as confirmed by subsequent detection
Arnaud Cassan 1, 9 10
L awrence Livermore National Labora- of a ~ 13 Earth-mass planet. Using a
Kem H. Cook 1,10 tory, Livermore, California, USA network of dedicated 1–2-m-class tele-
Christian Coutures 1, 3 11
University of Tasmania, School of scopes, we have entered a new phase
Stefan Dieters 1,11 Mathematics and Physics, Hobart, of planet discovery, and will be able to
Martin Dominik 1, 2,12,13 Tasmania, Australia provide constraints on the abundance
Dijana Dominis-Prester 1,14 12
Scottish Universities Physics Alliance, of frozen Super-Earths in the near future.
Jadzia Donatowicz 1,15 University of St Andrews, School of
Pascal Fouqué 1,16 Physics and Astronomy, St Andrews,
John Greenhill 1,11 United Kingdom The discovery of extrasolar planets is ar-
Kym Hill 1,11 13
Royal Society University Research guably the most exciting development
Matie Hoffman 1, 8 Fellow in astrophysics during the past decade,
Uffe G. Jørgensen 1,17 14
University of Rijeka, Physics Depart- rivalled only by the discovery of the cos-
Stephen Kane 1,18 ment, Faculty of Arts and Sciences, mic acceleration. The unexpected variety
Daniel Kubas 1, 6 Rijeka, Croatia of giant exoplanets, some very close
Jean-Baptiste Marquette 1, 3 15
Technische Universität Wien, Austria to their stars, many with high orbital ec-
Ralph Martin 1,19 16
Observatoire Midi-Pyrénées, Labora- centricity, has sparked a new generation
Pieter Meintjes 1, 8 toire d’Astrophysique, Université Paul of observers and theorists to address
John Menzies 1, 20 Sabatier, Toulouse, France the question of how planets form in the
Karen Pollard 1, 4 17
Niels Bohr Institutet, Astronomisk context of protostellar accretion discs.
Kailash Sahu 1, 21 Observatorium, København, Denmark Planets are now known to migrate and
Christian Vinter 1,17 18
Department of Astronomy, University of maybe even be ejected, via planet-disc
Joachim Wambsganss 1, 9 Florida, Gainesville, Florida, USA and planet-planet interactions. We are
Andrew Williams 1,19 19
Perth Observatory, Perth, Australia beginning to discover how our Solar Sys-
Kristian Woller 1,17 20
South African Astronomical Observa- tem fits into a broader community of
Marta Zub1, 9 tory planetary systems, many with very differ-
Keith Horne 1, 2,12 21
Space Telescope Science Institute, ent properties. Microlensing-based
Alasdair Allan 2, 22 Baltimore, Maryland, USA searches play a critical role by probing for
Mike Bode 2, 23 22
eSTAR Project, School of Physics, cool planets with masses down to that
Daniel M. Bramich 1, 2, 24 University of Exeter, Devon, United of Earth. Of key interest is how planets
Martin Burgdorf 2, 23 Kingdom are distributed according to mass and or-
Stephen Fraser 2, 23 23
Astrophysics Research Institute, bital distance (Figure 1) as this information
Chris Mottram 2, 23 Liverpool John Moores University, provides a crucial test for theories of
Nicholas Rattenbury 2, 25 Birkenhead, United Kingdom planet formation. Core accretion models
Colin Snodgrass 2, 6 24
The Isaac Newton Group of Tele- (Ida and Lin, 2005) are today the best
Iain Steele 2, 23 scopes, Santa Cruz de La Palma, description for the formation of planetary
Yiannis Tsapras 2, 23 Canary Islands systems: the accretion of planetesimals
25
The University of Manchester, Jodrell leads to the formation of cores, which
Bank Observatory, Macclesfield, then start to accrete gas from the primi-
1
PLANET Collaboration Cheshire, United Kingdom tive nebula. This scenario predicts that for
2
RoboNet Collaboration M dwarf stars there is a preferential for-
3
Institut d’Astrophysique de Paris, mation of Earth- to Neptune-mass planets
CNRS, Université Pierre et Marie Curie, In order to obtain a census of planets in 1–10 AU orbits. These planets are ex-
Paris, France with masses in the range of Earth to pected to form within a few million years.
4
University of Canterbury, Department Jupiter, eight telescopes are being used More massive planet (Jupiter) formation
of Physics and Astronomy, by the combined microlensing cam- is hampered by a longer formation time
­Christchurch, New Zealand paign of the PLANET and RoboNet col- (10 Myr) during which the gas evaporates
5
University of Notre Dame, Department laborations for high-cadence photo- and is no longer available to be accreted.
of Physics, Notre Dame, Indiana, USA metric round-the-clock follow-up of
6
ESO ongoing events, alerted by the OGLE There is a wide variety of planets and at
7
McDonald Observatory, Fort Davis, and MOA surveys. In 2005 we detected first sight it appears that our system
Texas, USA a planet of 5.5 Earth masses at 2.6 AU is very special. However, our view of the
from its parent 0.22 MA M star. This whole picture is still blurred by obser-
object is the first member of a new class vational biases inherent to the detection

The Messenger 128 – June 2007 33

Astronomical Science Beaulieu J.-P. et al., Hunting for Frozen Super-Earths via Microlensing

techniques using transits and radial ve- Known exoplanets: 11 + 180 + 4 = 195 (January 2007)
locities. Both methods are more sensitive Orbit period in years (for Solar-mass star)
to massive planets close to their parent 3d 30 d 1yr 10 yr 100 yr
star. Doppler measurements and the Doppler: 180
104
space transit missions, such as COROT,
can already, or will shortly, be able to de-

UN
µF
tect Neptune-mass planets close to their 1000

et,
parent star. Direct detections fill the other J

oN
extreme of very large separations which

ob
Planet mass in Earths

100 S

/R
are unknown in our Solar System. It is

ET
Kec
k

AN
therefore necessary to use different tech- , VL U N
10 TI

PL
niques, each probing different areas of AT
ASP, H

,
SIM

OA
the planet-mass versus orbital distance TrES, W
OGLE,

,M
parameter space. 1 V E

PF
LE

M
OG
Kepler
Already with ground-based observations, 0.1 M
the microlensing technique is sensitive M
to cool planets with masses down to that 0.01 Transits: 11 (+ 3) Microlensing: 4
of the Earth orbiting 0.1–1 MA stars, the Astrometry: 0
most common stars of our Galaxy, in or- P
bits of 1–10 AU. Currently, over 700 micro- 0.01 0.1 1 10 100
lensing events towards the Galactic Bulge Orbit radius in Astronomical Units (AU)
are alerted in real time by the OGLE and
MOA surveys each year. During these Figure 1 (above): Exoplanet discovery space (planet Figure 2 (below): The left panel shows the caustic
mass versus orbit size) showing the 8 planets from structure of a star/planet lens, with two possible tra-
events, a source star is temporarily mag-
our Solar System (labelled as letters), 180 Doppler jectories of a source star. The right panel shows
nified by the gravitational potential of wobble planets (black triangles), 11 transit planets the corresponding observed light curves. Hitting the
an intervening lens star passing near the (blue circled crosses), and 4 microlensing planets planetary caustic, or passing close to it, induces a
line of sight, with an impact parameter (red circled crosses). Also outlined are the regions short-lived but clearly detectable photometric signal.
that can be probed by different methods: Doppler,
smaller than the Einstein ring radius RE, a
transits; astrometry; and microlensing from the
quantity which depends on the mass of ground and space. Microlensing is a cost-efficient
the lens, and the geometry of the align- way to measure the mass function of cool planets
ment. For a source star in the Bulge, with down to the mass of the Earth.
a 0.3 MA lens, RE ~ 2 AU, the projected
angular Einstein ring radius is ~ 1 mas, and Impact angle = 271.0 (deg)
0.10 Mass ratio q = 0.002
the time to transit RE is typically 20–30 Separation d = 1.2
Impact angle = 50.0 (deg)
17.5
days, but can be in the range 5–100 days.
0.05

A planet orbiting the lens star generates a

I-Band magnitude

Planetary caustic
caustic structure in the source plane, with Central caustic 18.0
θ2 /θE

0.00
one small caustic around the centre of
mass of the system, the central caustic,
and one or two larger caustics further – 0.05
18.5

away, the planetary caustics. If the source

star happens to reach the vicinity of one – 0.10
of the caustics, its magnification is signifi- 2 520 2 540 2 560 2 580
– 0.10 – 0.05 0.00 0.05 0.10
cantly altered as compared to a single θ1 /θE
HJD - 2.450.000

lens, resulting in a brief peak or dip in the

observed light curve.
The duration of such planetary lensing
anomalies scales with the square root of
the planet’s mass, lasting typically a
few hours (for an Earth) to 2–3 days (for a
Jupiter). These two caustics (Figure 2)
provide two modes for detection. When
the central caustic is approached for all
events with a small impact angle between
source and lens star, corresponding to
a large peak magnification of the event,
the detection of planets in such events space to derive the planet/star mass ratio
becomes highly efficient (Griest and q, and the projected separation d in units
Safizadeh 1998). In contrast, planetary of RE. In general, model distributions for
caustics are only approached for a spe- the spatial mass density of the Milky Way,
cific range of orientations of the source the velocities of potential lens and source
trajectory, but the characterisation of a stars, and a mass function of the lens
planetary signal is much easier for such stars are required in order to derive prob-
configurations. ability distributions for the masses of the
planet and the lens star, their distance, as
The inverse problem, finding the prop- well as the orbital radius and period of
erties of the lensing system, is a complex the planet, by means of Bayesian analysis.
nonlinear one within a wide parameter

34 The Messenger 128 – June 2007

The observational challenge is to monitor
ongoing microlensing events, detected
by the OGLE and MOA survey telescopes,
with a fleet of telescopes to achieve
round-the-clock monitoring and detect
real-time deviations in the photometric
signal. The telescopes belonging to our
network together with their locations are
shown in Figure 3. During the coming
three Galactic Bulge seasons (from May
to September 2007, 2008, 2009 in the
southern hemisphere) we are planning to
use the eight telescopes of the PLANET/
RoboNET networks: Danish 1.5-m at
La Silla (Chile), Canopus 1.0-m at Hobart
and Bickley 0.6-m at Perth (Australia),
Rockefeller 1.5-m at Bloemfontein and
SAAO 1.0-m at Sutherland (South Africa).
These are the standard telescopes of the
PLANET network, to which were added
in 2004 two robotic telescopes of the
UK RoboNet network, North Faulkes 2-m
in Hawaii and Liverpool 2-m in Canary is-
lands, joined in 2006 by the South Faulkes
2-m in Australia.
Observing strategy, and description of
the reduction pipelines
A typical observing season of the Galac-
tic Bulge starts at the beginning of May
every year and lasts four months. Among
the 691 alerts available in 2006 (579 from
OGLE-III and 112 additional from MOA-II),
about 180 are available every night in the
middle of the season. Of these, around
20 targets can be monitored by 1-m-class
telescopes, whereas the Danish 1.54-m
and the 2-m telescopes can follow more
events. Therefore, we must apply some
criteria to select our 20 targets for every
observing night. This is done by one
member of the collaboration acting as a
coordinator, the so-called ‘homebase’.
Depending upon the current magnifica-
tion, the source brightness, and the
time of the last observation, a priority al-
gorithm assigns a worth to each of the
events and suggests sampling rates, with
the goal to maximise the planet detection
efficiency. If the magnification of one
event becomes very high, it may become
the sole designated target during that
night. While these suggestions are direct-
ly submitted to intelligent agents steering
the robotic telescopes of the RoboNet
network, the homebase currently tunes
them using our experience gained, before
In the case of OGLE, which has accumu-
Liverpool Telescope 2.0-m
lated many images of a given field before
a microlensing event is detected there,
SAAO 1.0-m the template is built from a set of the best
Rockefeller 1.5-m
images and is not held fixed throughout
the season. But in our case, we start ob-
serving an event when receiving the
Danish 1.5-m OGLE or MOA alert, so we have to build
the template ‘on the fly’. This generates
Perth 0.6-m
problems when new images appear after
Faulkes North 2.0-m a few nights, which are better than the
Canopus 1.0-m
first template. We then have to re-run the
process routine on all images of the event.
Faulkes South 2.0-m
A different image subtraction pipeline
is used on the RoboNet telescopes, but it
follows the same philosophy.
Figure 3: The different telescopes of The typical uncertainty of the on-line
the PLANET/RoboNet network. photometry is 1.2 % for an I = 17.8 mag
Galactic Bulge star at the Danish 1.5-m
instructing observers at the PLANET tele- telescope, and allows an on-line detec-
scopes by means of a web page. We tion of a deviating signal. When this ap-
plan to embed our experience into future pears, excitement grows and an alert to
advanced versions of the priority algo- homebase is issued. Homebase then
rithm and further automate this process. prompts an off-line reduction of the event
images, which are regularly uploaded to
At the beginning of the night, the observ- the Paris central archive. The off-line re-
er finds on the PLANET web pages the duction is done with our other image sub-
list of targets with sampling intervals set traction pipeline, pySIS, which facilitates
up by the homebase. He then defines the ‘fine-tuning’, so as to get the best possi-
exposure times for each target and re- ble photometry but is more difficult to
ports them on our private web page, so automate for real-time use. If the off-line
that the homebase can estimate the ob- reduction confirms the deviation, an alert
serving load at each telescope. Typical is issued to the microlensing community
sampling intervals are 0.5, 1, and 2 hours, to intensify observations and maximise
according to the priority of each event. the chances of a good characterisation
However, in case of a high magnification of the deviation, which is absolutely nec-
event, when the sensitivity to a planet is essary for future modelling of the event.
maximal, the sampling interval can be re- Moreover, all photometric data are made
duced to a few minutes, to the exclusion public immediately, as assistance to
of all other candidate objects. all teams in order to maximise the planet
hunting community’s success.
At the end of the exposure, the image is
pre-processed (bias, dark removed and
flatfielded), gets a standard name and The discovery of the frozen superEarth
is passed to an on-line pipeline. Starting OGLE-2005-BLG-390Lb
in 2006, on all the PLANET telescopes,
we shifted from a DoPhot-based on-line On 11 July 2005, the OGLE Early Warning
pipeline to an image subtraction pipeline System announced the microlensing
based on ISIS (Alard 2000). This robust event OGLE 2005-BLG-390, with a rela-
implementation, named WISIS, has two tively bright G4III giant as the source star.
main tasks: process takes all available PLANET/RoboNet included it in its list
images of a given event, chooses the best of targets and started to monitor it on
template and subtracts all images from 25 July. The microlens peaked at a mag-
that template after convolving the refer- nification Amax = 3 on 31 July. We were
ence point spread function (PSF) with the planning to continue to monitor it until the
kernel to mimic the current PSF. The source exited the Einstein ring, when on
update task only processes new images 10 August observers at the Danish tele-
using a previously chosen template. scope noticed a measurement deviating
The Messenger 128 – June 2007 35
Astronomical Science
by 0.06 mag from the point source point
lens prediction. They then took a second
measurement, deviating by 0.12 mag.
OGLE data became available, confirming
the deviation seen in Chile. In order to
check the nature of the deviation, home-
base increased the proposed sampling
Magnification
rate at the automated Perth telescope.
Perth started to observe this event con-
tinuously as soon as the target was within
reach. South Africa was clouded out,
and when observations resumed in Chile,
it was clear that the anomaly was over.
Different telescopes continued to ob-
serve the microlensing event. Perth data
– which were received only with some
delay – finally confirmed the short-dura-
tion deviation with a good coverage of six
additional data points. Combined with
two additional independent data points
from the MOA team (Mt. John, New Zea-
land), the evidence of a well-covered
short-term deviation from a point-lens
light curve was on record (see Figure 4).
Frenetic modelling activities started and it
became clear very quickly that we had
discovered a low-mass planet. The analy-
sis has proven to be rather straightfor-
ward for this event involving the transit of
a large source star over a planetary caus-
tic (Figure 5). The modelling of the photo-
metric data yields the mass ratio q = 7.6
± 0.7 10 –5, and the projected planet sepa-
ration d = 1.61 ± 0.008 (in units of RE,
the Einstein ring radius). We performed a
Bayesian analysis using Galactic mod-
els and a mass function in order to derive
probability distributions for the lens pa-
rameters (see Figure 2 from Beaulieu et
al. 2006) and a constraint on the nature
of the lens (low-mass main-sequence star
or stellar remnant). The median values
yield a host star of mass 0.22 +0.21
–0.11 M A lo-
cated at a distance of 6.6 ± 1.1 kpc within
the Galactic Bulge, orbited by a 5.5 +5.5
–2.7
Earth mass planet at an orbital separation
of 2.6 +1.5
–0.6 AU (Figure 6, and the artist
view of the planet by Herbert Zodet in
ESO Press Release 02/06).
Planet detection efficiency
The detection efficiency of the experi-
ment can be determined from all collect-
ed data, and comparison with the detec-
tions (or the absence of such) allows
conclusions about the planet abundance
36 The Messenger 128 – June 2007
Beaulieu J.-P. et al., Hunting for Frozen Super-Earths via Microlensing
3 3
OGLE
2
2.5
1 1.3
2 –1 0 9.5
OGLE
RoboNet
1.5
Canopus
1
– 20 0 20
Days since 31 July 2005 UT
Figure 4: Light curve of OGLE-2005-BLG-390Lb,
showing a brief planetary anomaly lasting for less
than a day observed by four telescopes. The lens
star is a ~ 0.22 M A Galactic Bulge M dwarf orbited
by a ~ 5.5 Earth-mass planet at ~ 2.6 AU with a
around the probed stars. The first at-
tempts were done on individual high-
magnification events using a point-source
approximation, and then were applied
to a sample of the 42 well-covered micro-
lensing events acquired by PLANET in
1995–1999 (Gaudi et al. 2002). Less than
1/3 of the lenses are orbited by Jupiters
with orbits in the range 1–5 AU. We are
currently working on an analysis combin-
ing 11 years of data (1995–2005). We
calculate the detection efficiency of each
microlensing light curve to lensing com-
panions as a function of the mass ratio
and projected separation of the two com-
ponents, now taking into account ex-
tended source effects. We use the same
Bayesian analysis as for determining
probability densities for the lens star and
planet properties (Dominik 2006). Fig-
ure 7 gives the mean detection efficiency
of PLANET combining 14 well-sampled
events from 2004. For Jupiter-mass
planets, the detection efficiency reaches
50 %, while it decreases only with the
square-root of the planet mass until the
detection of planets is further suppressed
by the finite size of the source stars for
planets with a few Earth masses. Never-
theless, the detection efficiency still re-
mains a few per cent for planets below
10 Earth masses, made of rock and ice.
As of today, four planets have been de-
1.6
1.5
1.4
Planetary
deviation
10 10.5 11
Danish
Perth
MOA
period of ~ 10 years (plotted as solid line). The best
alternative model is a binary source star but is re-
jected by the data. The dashed line is the point source
point lens model without the planetary deviation
(Beaulieu et al. 2006).
tected by microlensing, two of which
are Jupiter analogues (Bond et al. 2004,
Udalski et al. 2005) and one Neptune
(Gould et al. 2006), where the perturba-
tion is due to the central caustic, and
one rocky/icy planet of 5.5 Earth mass
named OGLE-2005-BLG-390Lb (Beau-
lieu et al., 2006) via a planetary caustic.
They are overplotted on Figure 7. So the
era of discovery of frozen Super-Earths
has been opened by the microlensing
technique. One of the consequences of
the discovery of such a small planet, is
that these small rock/ice planets should
be common.
Obtaining more information about these
planets
Unlike other techniques, microlensing
does not offer much chance to study the
planetary system in more detail because
the phenomenon only occurs once for
each star. Only a significant statistical
sample will allow us to reach firm conclu-
sions and finally answer the question of
how special is our own Solar System. Ad-
ditional information about a specific event
can be obtained once the lens star is di-
rectly detected. Here, we must wait many
years till the relative motion of the source
and lens stars separate them on the sky.
 Figure 6: Perturbation of the image of the source
star by the planetary caustic observed with an
ideal telescope. The time interval between the
images is 0.1 day. The image scale is 30 microarcsec
(Bennett and Williams).

0.10

0.05
θ2 /θE

0.00

– 0.05

– 0.10

0.90 0.95 1.00 1.05 1.10

θ1 /θE

Figure 5: The source star and the planetary caustic

of OGLE-2005-BLG6390. Notice the small size of
the planetary caustic compared to the source star.

%
50

%
25
Bennett et al. (2006) using HST images 1000
have detected the lens star in the micro-
lensing event OGLE-2003-BLG-235/
Planet mass (M Earth )

25

MOA-2003-BLG-53, and therefore the

%

uncertainty on the planetary parameters

5%

have been greatly reduced. This could 100

5%

be achieved too with HST or adaptive op-

tics for OGLE-2005-BLG-169. In the case
2%

of the lens OGLE-2005-BLG-390La, the

observation is much more difficult since
2%

we would need to detect a K ~ 22 mag 10

object at about 40 mas (in five years) from
a star that is 10 mag brighter. In the com-
ing years, statistics about frozen Super-
Earth planets orbiting M and K dwarfs will
be obtained, complementing the param-
eter space explored by space transit mis-
sions like COROT and KEPLER or ag-
gressive ground-based Doppler search,
like those using the CORALIE, SOPHIE
and HARPS instruments.
0.1
Semi-major axis (AU)
is supported by the Agence Nationale de La Recher-
che. KHC work performed under auspices of US
DOE, by UC, LLNL under contract, W 7405-Eng-48.
NJR acknowledges financial support by a PPARC
PDRA fellowship and is partially supported by
the European Community’s Sixth Framework Marie
Curie Research Training Network Programme, Con-
tract No. MRTN-CT-2004-505183 ANGLES.
1.0 10.0
Figure 7: Average detection efficiency of planets as a
function of mass and orbital separation (assuming
circular orbits) in the 14 favourable events monitored
by PLANET in 2004 (preliminary analysis) along
with the planets detected by microlensing as of 2006
marked as points.

Acknowledgements
Microlensing follow-up observations would not hap-
pen without the alert systems of the OGLE and MOA
collaborations. We are very grateful to OGLE and
MOA, and to the observatories that support our sci-
ence (ESO, Canopus, CTIO, Perth, SAAO, Boyden)
via the generous allocations of time that make this
work possible. The operation of Canopus Obser-
vatory is in part supported by a financial contribution
from David Warren, and the Danish telescope at
La Silla is operated by IDA financed by SNF. The
RoboNet project is funded by the UK STFC. HOLMES
References
Alard C. 2000, A&AS 144, 363
Beaulieu J. P. et al. 2006, Nature 439, 437
Bennett D. P. et al. 2006, ApJ 647, 171
Bond I. A. et al. 2005, ApJ 606, 155
Dominik M. 2006, MNRAS 367, 669
Gaudi B. S. et al. 2002, ApJ 566, 463
Gould A. et al. 2006, ApJ 644, L 37
Griest K. and Safizadeh N. 1998, ApJ 500, 37
Ida S. and Lin D. N. C. 2005, ApJ 626, 1045
Udalski A. et al. 2005, ApJ 628, 109
Relevant websites
PLANET: http://planet.iap.fr/
RoboNet: http://www.astro.livjm.ac.uk/RoboNet/
MicroFUN: http://www.astronomy.ohio-state.
edu/~microfun/
MOA website: http://www.phys.canterbury.ac.nz/
moa/
OGLE website: http://www.astrouw.edu.pl/~ogle/

The Messenger 128 – June 2007 37

Astronomical Science
Sulphur Abundances in Metal-Poor Stars –
First Result from CRIRES Science Verification
Poul Erik Nissen 1
Martin Asplund 2
Damian Fabbian 2
Florian Kerber 3
Hans Ulrich Käufl 3
Max Pettini 4
1
Department of Physics and Astronomy,
Aarhus University, Denmark
2
Research School of Astronomy and
Astrophysics, Australian National Uni-
versity
3
ESO
4
Institute of Astronomy, University of
Cambridge, United Kingdom
Sulphur is the tenth most abundant ele-
ment in the Universe and plays an im-
portant role in studies of the chemical
enrichment and star-formation history
of distant galaxies. Due to the lack of
suitable sulphur lines in the visible part
of stellar spectra there is, however, still
no agreement on the abundance of sul-
phur in Galactic metal-poor stars, and
we are therefore uncertain about the
nucleosynthetic origin of sulphur. New
observations of infrared sulphur lines
with the cryogenic high-resolution infra-
red echelle spectrograph (CRIRES)
at ESO’s VLT are helping to solve this
problem.
Abundance ratios of elements in celestial
objects are usually given on a logarithmic
scale with the corresponding solar ratio
as the zero point. From the ratio between
the number of iron and hydrogen atoms,
we define the metallicity of a star as [Fe/H]
≡ log(NFe /NH) Star – log(NFe /NH) Sun, and in
order to get information on the nucleo-
synthesis of an element X, we are study-
ing how the quantity [X/Fe] ≡ log(NX /NFe) Star
– log(NX /NFe) Sun varies as a function of
[Fe/H]. For example, one finds that mag-
nesium has a constant overabundance
[Mg/Fe] = + 0.3 dex (corresponding to a
factor of two) in Galactic halo stars with
metallicities in the range – 4 < [Fe/H] < –1,
Relative flux
whereas [Mg/Fe] declines continuously to
zero in disc stars with –1 < [Fe/H] ≤ 0. A
similar trend is found for [Si/Fe]. This can
be explained if Mg and Si are made pri-
marily by a-capture reactions (i.e. suc-
cessive captures of a-particles) in Type II
supernovae (SNe), whereas iron is made
38 The Messenger 128 – June 2007
by silicon burning in both Type Ia and
Type II SNe. The point is that Type Ia SNe
are not contributing with iron until the
disc phase of our Galaxy, because their
occurrence is delayed by about one
billion years relative to Type II SNe, due
to the lower progenitor mass of Type Ia
SNe.
Sulphur is an element with an even num-
ber of protons (Z = 16) like Mg and Si
(Z = 12 and 14), and nucleosynthesis cal-
culations predict that S is made in the
same way as Mg and Si in Type II SNe.
Hence, we expect [S/Fe] to have the
same trend with [Fe/H] as [Mg/Fe] and
[Si/Fe]. Sulphur abundances derived from
the weak S i line at 869.5 nm by Israelian
and Rebolo (2001) and Takada-Hidai et
al. (2002) showed, however, an increasing
trend of [S/Fe] towards low metallicities,
with [S/Fe] reportedly reaching values as
high as + 0.8 dex (a factor of six higher
than the S/Fe ratio in the Sun) at [Fe/H] =
– 2.0. They suggested that such high
S/Fe ratios might be due to an enhanced
sulphur production in supernovae with
a very large explosion energy, so-called
hypernovae. Nissen et al. (2004), on the
other hand, used the stronger S i lines
at 921.3 and 923.8 nm to derive a near-
constant [S/Fe] ~ + 0.3 for halo stars in
the metallicity range – 3 < [Fe/H] < –1, as
expected if normal Type II SNe are the
sole source of sulphur. More recently,
Caffau et al. (2005) proposed a dichot-
omy of [S/Fe] among Galactic halo stars
with both high and low [S/Fe] values,
which would imply a very complicated
evolution of the sulphur abundance in our
Galaxy.
A reason for the diverging [S/Fe] results
may be that errors in determining sul-
phur abundances are larger than claimed
by the authors of the cited papers. In this
1.05
1.00
0.95
0.90
0.85
0.80
1.0454 1.0455 1.0456 1.0457 1.0458
Wavelength (µm)
connection, we note that the 869.5 nm
S i line is very weak in metal-poor halo
stars and hence the measured strength
of the line may be affected by irregular
fringing variations of the CCD detector
response, which are difficult to correct by
flat-fielding. The stronger S i lines at 921.3
and 923.8 nm are less affected by such
errors, but they occur in a spectral region
hampered by numerous telluric lines that
are often blended with the sulphur lines.
CRIRES observations
CRIRES is a cryogenic, infrared echelle
spectrograph designed to provide a re-
solving power l/∆l of up to 100 000 be-
tween about 950 nm and 5 000 nm as
described in detail by Käufl et al. (2006).
The commissioning of this VLT instru-
ment opens up a new opportunity for in-
dependent determinations of sulphur
abundances in metal-poor stars based
on high-resolution observations of the
S i triplet at 1.046 µm. As part of the sci-
ence verification of CRIRES, a spectrum
around the infrared triplet was obtained
for the halo dwarf star G29-23 (V = 10.19,
[Fe/H] = –1.7) on 6 October 2006. The
entrance slit width of CRIRES was set at
0.4 arcsec, which corresponds to a re-
solving power of 50 000 with four detec-
tor pixels per spectral resolution bin ∆l.
In order to improve the removal of sky
emission and detector dark current, the
observations were performed in nod-
ding mode with a shift of 10 arcsec be-
tween the two settings of the star on the
slit. The exposure time was 2 400 s. The
seeing was rather poor (about 1.3 arcsec)
but adaptive optics was applied to im-
prove the stellar image, and the com-
bined spectrum has a very satisfactory
signal-to-noise ratio S/N ~ 330 per spec-
tral dispersion pixel. This is considerably
Figure 1: The CRIRES spectrum of the
metal-poor ([Fe/H] = –1.7) dwarf star
G29-23 around the 1.046 µm S i triplet
(dots) compared with synthetic model-
atmosphere line profiles for three
sulphur abundances corresponding to
1.0459 1.0460 1.0461 [S/Fe] = 0.0, +0.3 and +0.6, respec-
tively.
better than our earlier UVES spectrum 1.05 Figure 2: The UVES spectrum of G29-
23 around the S i lines at 921.3, 923.8,
of the same star (observed under similar 1.00
and 869.5 nm (dots) compared with
conditions and with the same exposure 0.95
synthetic line profiles for three sulphur

Relative flux
time), which has S/N ~ 200 around the 0.90 abundances corresponding to [S/Fe] =
S i lines at 921.3 and 923.8 nm. Further- 0.85 0.0, +0.3 and +0.6, respectively.
more, unlike the UVES near-IR spectrum, 0.80
the CRIRES spectrum at 1.046 µm is not 0.75
plagued by telluric lines and fringing resi- 0.70
9 211.5 9 212.0 9 212.5 9 213.0 9 213.5 9 214.0 9 214.5
duals.
1.04
The CRIRES spectrum of G29-23 is 1.00
shown in Figure 1 in comparison with
Relative flux

0.96
synthetic profiles of the sulphur lines
0.92
for three values of [S/Fe]. Details of the
model-atmosphere calculations and 0.88
the determination of the iron abundance, 0.84
which is based on Fe ii lines, can be found 0.80
in Nissen et al. (2007). As can be seen 9 236.0 9 236.5 9 237.0 9 237.5 9 238.0 9 238.5 9 239.0

from the Figure, the synthetic profile cor- 1.02

responding to [S/Fe] = + 0.3 provides 1.01
an excellent fit to the CRIRES data for all 1.00
Relative flux

three S i lines. 0.99

0.98
0.97
Comparison with UVES observations
0.96
0.95
In Figure 2, the UVES spectrum of G29- 8 693.0 8 693.5 8 694.0 8 694.5 8 695.0 8 695.5 8 696.0
23 is shown for the sulphur lines at 869.5, Wavelength (Å)

921.3 and 923.8 nm. Telluric lines were

removed by dividing by a scaled B-type
star spectrum as described in Nissen et
al. (2007). Again, the synthetic profile cor-
responding to [S/Fe] = + 0.3 fits the data 0.8 Figure 3: [S/Fe] versus [Fe/H] for
Galactic stars. Disc stars ([Fe/H] > –1)
very well, but it is clear that the [S/Fe] val-
0.6 from Chen et al. (2002) are shown
ue obtained from the weak 869.5 nm line with green crosses. For the halo stars
is rather uncertain. 0.4 ([Fe/H] < –1), blue circles with error
[S/Fe] (LTE)

bars show data based on sulphur

abundances derived from the 921.3,
Although one should not put too much 0.2
923.8 nm S i lines, and red circles
weight on a single star, the CRIRES spec- show data from the weak 869.5 nm
0.0
trum of G29-23 provides an important S i line. In the upper panel LTE has
check of the sulphur abundances ob- – 0.2 been assumed in deriving the S abun-
dances, whereas the lower panel
tained from UVES spectra for 40 halo
includes non-LTE corrections from
stars by Nissen et al. (2007). As shown in – 0.4
– 3.5 – 3.0 – 2.5 – 2.0 –1.5 –1.0 – 0.5 0.0 0.5 Takeda et al. (2005).
the upper panel of Figure 3, the UVES
0.8
data indicate that [S/Fe] lies around a pla-
teau of + 0.3 dex if local thermodynamic 0.6
equilibrium (LTE, i.e. a Boltzmann distribu-
tion of the atoms over the possible atomic
[S/Fe] (non-LTE)

0.4
energy states) is assumed when deriving
0.2
the S abundances. If non-LTE corrections
based on the statistical equilibrium cal- 0.0
culations of Takeda et al. (2005) are tak-
en into account, the plateau decreases to – 0.2
about + 0.2 dex as shown in the lower
– 0.4
panel of Figure 3. – 3.5 – 3.0 – 2.5 – 2.0 –1.5 –1.0 – 0.5 0.0 0.5
[Fe/H]

The scatter of [S/Fe] around this plateau

is remarkably small, ± 0.07 dex only,
and no star has an enhanced S/Fe ratio

The Messenger 128 – June 2007 39

Astronomical Science Nissen P. E. et al., Sulphur Abundances in Metal-Poor Stars

([S/Fe] > + 0.60) as claimed by Israelian
and Rebolo (2001), Takada-Hidai et al.
(2002) and Caffau et al. (2005). Our re-
sults suggest that sulphur in the Galactic
halo was made in the same way as
Mg and Si, i.e. by a-capture processes in
massive SNe.
The S/Zn ratio of Damped Lyman-alpha
systems
The sulphur abundances derived by
Nissen et al. (2007) were combined with
zinc abundances determined from the
472.2 and 481.1 nm Zn i lines. Both S and
Zn are among the few elements which
are not readily depleted onto dust in the
interstellar medium. For this reason,
they are key to studies of metal enrich-
ment in distant galaxies, particularly
those detected as damped Lyman-alpha
systems (DLAs) in the spectra of high-
redshift quasars. Assuming that sulphur
behaves like other a-capture elements
and that Zn follows Fe, the S/Zn ratio has
been used to estimate the timescale
of the star-formation process in DLAs. In
particular, it has been argued that a solar
S/Zn ratio indicates that Type Ia SNe
have contributed to the chemical enrich-
ment and that the age of a DLA system
with a solar-like S/Zn therefore must be at
least one billion years.
Before such a conclusion can be made,
it is important to clarify the trend of [S/Zn]
[S/Zn]
there is an overall decrease in [S/Zn] at all
but the highest values of [Zn/H] consid-
ered here. Furthermore, non-LTE effects
are most significant at low metallicities
with the result that, apparently, [S/Zn] re-
verts to solar values when [Zn/H] < – 2.
Such behaviour is unusual but, given our
current limited understanding of the nu-
cleosynthesis of Zn, cannot be excluded.
Taken at face value, the lack of a strong
metallicity trend in the lower panel of
Figure 4 would indicate that the useful-
ness of the S/Zn ratio as a ‘clock’ of
the star-formation history is rather limited.
The question that remains concerns
the accuracy of the non-LTE calculations
by Takeda et al. (2005); they depend
critically on the rather uncertain cross-
section for inelastic collisions with neutral
hydrogen atoms, which tend to enforce
an LTE population of the energy levels
(Asplund 2005). Future observations of
the forbidden sulphur line at 1.082 µm in
0.6
0.4
0.2
0.0
cool halo giants may be particularly help-
ful in this connection. This line is formed
close to LTE in cool stars, because near-
ly all sulphur atoms are in the lower ener-
gy level of the line (the ground state of
the S i atom), and the number of atoms in
the upper level is determined by colli-
sions rather than radiative transitions. The
forbidden line is very weak, but thanks
to the efficiency and high resolution of
CRIRES, precision measurements will be
possible at metallicities as low as [Fe/H] ~
– 2.5.
References
Asplund M. 2005, ARA&A 43, 481
Caffau E. et al. 2005, A&A 441, 533
Chen Y. Q. et al. 2002, A&A 390, 225
Israelian G. and Rebolo R. 2001, ApJ 557, L43
Käufl et al. 2006, The Messenger 126, 32
Nissen P. E. et al. 2004, A&A 415, 993
Nissen P. E. et al. 2007, A&A 469, 319
Takada-Hidai M. et al. 2002, ApJ 573, 614
Takeda Y. et al. 2005, PASJ 57, 751
Figure 4: [S/Zn] versus [Zn/H] for Gal-
actic stars and damped Lyman-alpha
systems. Typical 1-s error bars are
shown. In the upper panel the stellar
S and Zn abundances have been de-
rived assuming LTE, whereas the

lower panel includes non-LTE correc-

among Galactic stars. Figure 4 shows
– 0.2 tions from Takeda et al. (2005).
the results from Nissen et al. (2007). It is
found that the trend of [S/Zn] depends – 0.4
quite critically on whether LTE is as- – 0.6
Stars LTE DLAs
sumed or not. When the non-LTE correc-
tions of Takeda et al. (2005) are applied, – 0.8
– 2.8 – 2.4 – 2.0 –1.6 –1.2 – 0.8 – 0.4 0.0 0.4

0.6

0.4

0.2

0.0
[S/Zn]

– 0.2

– 0.4
Stars, non-LTE DLAs
– 0.6

– 0.8
– 2.8 – 2.4 – 2.0 –1.6 –1.2 – 0.8 – 0.4 0.0 0.4
[Zn/H]

40 The Messenger 128 – June 2007

Astronomical Science
Using Globular Clusters to Test Gravity in the
Weak Acceleration Regime
Riccardo Scarpa 1
Gianni Marconi 2
Roberto Gilmozzi 2
Giovanni Carraro 3
1
Instituto de Astrofísica de Canarias,
La Laguna, Tenerife, Spain
2
ESO
3
Università di Padova, Italy
We report on the results from an on-
going programme aimed at testing
Newton’s law of gravity in the low ac-
celeration regime using globular clus-
ters. We find that all clusters studied so
far behave like galaxies, that is, their
velocity dispersion profiles flatten out at
large radii where the acceleration of
gravity goes below 10 –8 cm s –2, instead
of following the expected Keplerian
fall-off. In galaxies this behaviour is as-
cribed to the existence of a dark-matter
halo. Globular clusters, however, are
not supposed to contain dark matter,
hence this result might indicate that our
present understanding of gravity in the
weak regime of accelerations is incom-
plete and possibly incorrect.
Newtonian dynamics and globular
clusters
Stars within galaxies, and galaxies within
clusters of galaxies, are very far apart
from each other. As a consequence, the
typical accelerations governing the
dynamics of galaxies are orders of mag-
nitude smaller than the ones probed
in our earth-based laboratories or in the
Solar System. Thus, any time Newton’s
law is applied to galaxies (e.g., to infer
the existence of dark matter), its validity is
severely extrapolated. Although there
are in principle no reasons to distrust
Velocity dispersion (km/s)
Newton’s law in the weak acceleration
regime, agreement has been reached
(e.g., Binney 2004) that galaxies start to
deviate from Newtonian dynamics, and
that dark matter is needed to reconcile
observations with predictions, always for
the same value of the internal accelera-
tion of gravity of a0 ~ 10 – 8 cm s –2. This
systematic property, more than anything
else, raises the possibility that we may
be facing a breakdown of Newton’s law,
rather than the effects of dark matter.
Since globular clusters are free falling to-
ward the Milky Way, their internal dy-
namics are only affected by tidal stress,
which is in most cases well below a0.
Therefore the internal dynamics of globu-
lar clusters can be used to probe the
same range of accelerations typical of
galaxies, without the complication of dark
matter. Following this idea, we have
studied the dynamics of the external re-
gions of w Centauri (Scarpa, Marconi
and Gilmozzi 2003). This massive cluster
was selected because proper motions
for several thousand stars were available
in the literature. Combining proper mo-
tions with radial velocity information al-
lows all three components of the velocity
vector to be obtained, thus fully address-
ing the possible effects of anisotropy. The
result is clearly shown by Figure 1. The
velocity dispersion profiles, as derived for
the three components of motion, are very
similar, showing that the cluster is iso-
tropic (as is also indicated, by the way,
from its very nearly circular shape). More-
over, the dispersion is found to be con-
stant at large radii.
Following this initial result, that could be
due to some other effect like tidal
heating, we collected data for three more
globular clusters: NGC 7078 (M15)
and NGC 6171 (Scarpa, Marconi, and
Gilmozzi 2004A,B), and NGC 7099
(Scarpa et al. 2006). Data for a fourth
cluster, NGC 6341 (M92), appeared
recently in the literature (Drukier et al.
2006) and this cluster is also presented
here. In all cases (see Figure 2) the
30
20
10
0 data. It is a Gaussian plus a constant drawn to
0 10 20
Distance from cluster centre (pc)
velocity dispersion profile mimics what is
observed in high surface brightness
elliptical galaxies. That is, the dispersion
is maximal at the centre, then decreases
toward a constant value at large radii,
where the acceleration goes below a0.
With data for five globular clusters, one
can start comparing results and, interest-
ingly, we found that in all cases the flat-
tening of the dispersion profile occurs for
very similar values of the internal acceler-
ation of gravity. Values are collected in
Table 1, where we present, for each clus-
ter, its absolute V magnitude, the radius
where the flattening occurs, and the cor-
responding acceleration derived assum-
ing a mass-to-light ratio of one. Within er-
rors all profiles flatten out at a0. It is worth
pointing out that these five clusters are
different in size, mass, position in the halo
of the Milky Way and dynamical history.
Therefore there is no obvious reason why
they should conspire to have a similar
velocity dispersion profile at large radii.
Galaxies provide us with another pow-
erful tool to further disentangle non-
Newtonian effects from other more
classical phenomena like tidal heating, or
a combination of effects like the dis-
tribution of dark remnants plus tidal heat-
ing plus cluster evaporation and so on,
that might be responsible for increasing
the velocity dispersion in the outskirts.
High surface brightness galaxies (HSB)
and low surface brightness galaxies
(LSB) are known to behave differently.
The latter have a remarkably flat velocity
dispersion profile (e.g., Mateo 1997;
Wilkinson 2006), allegedly due to LSB
galaxies being dark matter dominated all
the way to their centre, while HSB galax-
ω Cen
Figure 1: The velocity dispersion profile of w Cen-
tauri. Circles and squares represent the disper-
sion as derived from proper motion data. Crosses
are radial velocity dispersions from the literature, to
which we added data for 75 stars (the four last
points with error bars). The solid line is not a fit to the
30 40 50 emphasise the flattening of the dispersion at large
radii.
The Messenger 128 – June 2007 41
 Astronomical Science Scarpa R. et al., Using GCs to Test Gravity in the Weak Acceleration Regime

12 Figure 2: Velocity dispersion profiles of the other four

clusters studied so far. In all cases the velocity dis-
NGC 7078 5 NGC 6171 persion is maximal at the centre, then decreases to
10 converge toward a constant value at large radii. This
is similar to the case of high surface brightness

Velocity dispersion (km/s)

Velocity dispersion (km/s)

4
elliptical galaxies. Note that the profiles of NGC 6171
8 and NGC 7099 were derived from our own data,
while the profiles of NGC 7078 and NGC 6341 were
3 derived from data in the literature (Drukier et al. 2006
6
and references therein). The solid line has the same
meaning as in Figure 1.
2
4

2 1

0 0
0 10 20 30 40 50 0 5 10 15 20
Distance from cluster centre (pc) Distance from cluster centre (pc)

figuration three 2700 s exposures were

NGC 7099 8 NGC 6341 obtained under good atmospheric condi-
6
tion (clear sky and seeing ~ 1 arcsec) on
29 and 30 August 2005.
Velocity dispersion (km/s)
Velocity dispersion (km/s)

4
2 small number of stars, to evaluate and
0 0
0 5
Distance from cluster centre (pc)
ies are baryon-dominated at the centre.
In view of what we have found for dense
globular clusters that probe the same
accelerations as HSB galaxies, it is natu-
ral to wonder whether low-concentration
globular clusters behave like LSB gal-
axies. That is, do low-concentration glob-
ular clusters have constant velocity dis-
persion?
The case of NGC 288
In an attempt to answer this question we
studied NGC 288, a low-concentration
cluster located at 8.3 kpc from the Sun
and 11.6 kpc from the Galactic Centre,
that has internal acceleration of gravity
everywhere below a0, as is the case for
LSB galaxies. The initial selection of
targets around NGC 288 was based on
colour, as derived from the analysis of
ESO Imaging Survey frames. A catalogue
of targets was prepared including mostly
stars from the sub-giant branch down
6
4 a template (the target with the best spec-
2 eliminate possible offsets in the velocity
10 15 20 0 5
Distance from cluster centre (pc)
to the turn off, between 15 and 18 appar-
ent V mag. Observations were then ob-
tained with FLAMES at the ESO VLT tele-
scope. FLAMES is a fibre multi-object
spectrograph, allowing the simultaneous
observation of up to 130 objects. We se-
lected the HR9B set-up that includes
the magnesium triplet covering the wave-
length range 5143 < l < 5 346 Å at res-
olution R = 25 900. Stellar astrometry was
derived cross correlating the stellar posi-
tions on the EIS frames with coordinates
from the US Naval Observatory (USNO)
catalogue, which proved to have the re-
quired accuracy (0.3 arcsec) for FLAMES
observations. Two different fibre configu-
rations were necessary to observe all the
selected stars (Figure 3). For each con-
Cluster Name
NGC 5139 (w Centauri)
NGC 6171 (M107)
NGC 6341 (M92)
NGC 7078 (M15)
NGC 7099 (M30)
Radial velocities were derived by cross-
correlating the spectra of each target with
trum). The two configurations shared a
zero point. A posteriori, we verified that
no correction was necessary down to a
level of accuracy of 250 m s –1, well below
10 15 20 the accuracy required for our study. Fi-
nally, keeping in mind that we are inter-
ested only in the velocity dispersion, the
global velocity zero point was derived
by identifying a few lines in the spectrum
of the template. Altogether 126 radial ve-
locities with accuracy better than 1 km s –1
were derived. Of these, all but two were
found to be cluster members, consistent
with the very low contamination expected
at the high Galactic latitude (b = – 89 de-
grees) of this cluster.
To better constrain the velocity disper-
sion close to the cluster centre, we com-
bined our data with data for 24 additional
stars, mostly within 6 pc from the clus-
ter centre, and radial velocity accuracy
better that 1 km/s (from Pryor et al. 1991).
After applying an offset of 2.9 km/s to
match our radial velocity zero point, these
Mv R (pc) a (cm s –2 ) Table 1: For each globular cluster
studied, the absolute magnitude,
–10.29 27 ± 3 2.1 ± 0.5 × 10 –8
radius at which the velocity dispersion
– 7.13 8±2 1.3 ± 0.6 × 10 –8
flattens and the corresponding ac­-
– 8.20 12 ± 2 1.5 ± 0.6 × 10 –8 celeration due to gravity for a mass-to-
– 9.17 20 ± 2 1.4 ± 0.4 × 10 –8 light ratio of 1, are listed.
– 7.43 10 ± 2 1.1 ± 0.4 × 10 –8

42 The Messenger 128 – June 2007

 – 35
Radial velocity (km/s)
– 40
– 45
– 50
0 5 10
Distance from cluster centre (pc)
Figure 3: Radial velocity distribution for the low-
concentration cluster NGC 288. The 124 stars stud-
ied as part of this work are shown as solid circles,
while 24 velocities from Pryor et al. 1991 are shown
as open squares. The data distributes uniformly
from the centre to 20 pc, showing no indication of a
decrease of the dispersion.
data smoothly merge with ours in the re-
gion of overlap, showing basically the
same velocity dispersion (Figure 3). This
combined sample was used to search
for ordered rotation in NGC 288 that might
contribute to sustain the cluster, finding
no evidence for rotation down to the level
of 0.5 km/s.
Velocities from this combined data set al-
lowed us to build a well-sampled velocity
dispersion profile from the centre to al-
most 18 pc (Figure 4). In Table 2 we re-
port the velocity dispersion in km/s with
1s uncertainties, together with the limits
of the bins, the number of stars in each
bin, and the bin centre, defined as the
mean of the radii of the stars in the bin.
Looking at both Figures 3 and 4, we see
no indications of a vanishing velocity dis-
persion at large radii, rather, the disper-
sion remains constant with an average
value of 2.3 ± 0.15 km/s over the full range
of radii covered by the data.
The five clusters w cen, NGC 7078,
NGC 6171, NGC 7099, and NGC 6341,
while having different sizes, different
masses, and different dynamical histo-
ries, have the common property of being
highly concentrated, thus the acceler-
ation of gravity in their central regions is
above a0. Only in the outskirts is the
acceleration below this value. All these
clusters are found to behave like HSB
elliptical galaxies, i.e. to have a constant
velocity dispersion at large radii. By
contrast, NGC 288 has the peculiar prop-
NGC 288 4 NGC 288
Velocity dispersion (km/s)
3
2
15 20 1
0
0 5
Distance from cluster centre (pc)
Bin limits (pc) Stars/bin
0–2 8 1.34
2–4 21
4–6 20
6–8 22
8 –10 20
10 –12 14
12–14 25
14–20 17
erty of being rather diffuse, with a central
surface brightness of ~ 20 mag arcsec –2
in the V band. This has the important
consequence that the internal accelera-
tion of gravity is extremely small. Indeed
it can be shown to be everywhere be-
low a0 for any of the typical mass-to-light
ratios assumed for globular clusters.
It is well known that LSB galaxies have
velocity dispersion profiles which are
remarkably flat, with no central maximum
(Mateo 1997; Wilkinson et al. 2006).
If this property is due to a breakdown of
Newtonian dynamics below a0 (and
not because of dark matter), then the ve-
locity dispersion of NGC 288 should
mimic what is observed in these galaxies.
Within the errors, this is certainly the case
(Figure 4). The similarity between NGC
288 and LSB galaxies is striking. The ca-
nonical explanation for the constant
velocity dispersion in LSB galaxies is that
these objects are dark matter domi-
nated all the way to their centres. In the
case of a globular cluster, this expla-
nation is quite unpalatable. Thus, in view
of these results for globular clusters and
also of the amazing ability of a particular
modification of the Newtonian Dynamics
Figure 4: The velocity dispersion pro-
file of the low-concentration cluster
NGC 288. From the point of view of
the internal acceleration of gravity, this
cluster is the equivalent of a low sur-
face brightness galaxy and, similarly to
what is found in these galaxies, the
dispersion profile is remarkably flat.
10 15 20
bin centre (pc) s (km s –1) Table 2: Data on the binned radial
2.32 ± 0.64 velocity dispersion for NGC 288.
3.21 2.41 ± 0.41
5.09 2.73 ± 0.46
7.18 2.52 ± 0.41
9.02 2.60 ± 0.45
10.71 2.16 ± 0.46
12.95 1.90 ± 0.31
16.12 2.17 ± 0.42
known as MOND (Milgrom 1983) to de-
scribe successfully the properties of a
large number of stellar systems without
invoking the existence of non-baryonic
dark matter, we see here evidence for a
failure of Newtonian Dynamics for ac-
celerations below a0. Given the potential
impact of this claim, we urge the astro-
nomical community to disprove or gen-
eralise our results.
References
Binney J. 2004, in “Dark Matter in Galaxies”,
ed. S. D. Ryder et al., IAUS 220, 3
Drukier G. A. et al. 2007, AJ 133, 1041
Mateo M. 1997, in “The nature of Elliptical Galaxies”,
ed. M. Arnaboldi, ASPC 116, 259
Milgrom M. 1983, ApJ 270, 365
Pryor C. et al. 1991, AJ 102, 1026
Scarpa R., Marconi G. and Gilmozzi R. 2003,
A&AL 405, 15
Scarpa R., Marconi G. and Gilmozzi R. 2004a,
IAUS 220E, 215
Scarpa R., Marconi G. and Gilmozzi R. 2004b, in
“Baryons in Dark Matter Halos”, eds. R. Dettmar
et al., SISSA, Proceedings of Science, 55.1,
http://pos.sissa.it
Scarpa R. et al. 2007, A&AL 462, 9
Wilkinson M. I. et al. 2006, The Messenger 124, 25
The Messenger 128 – June 2007 43
Astronomical Science
Dissecting the Nuclear Environment of Mrk 609
with SINFONI – the Starburst-AGN Connection
Jens Zuther 1
Sebastian Fischer 1
Jörg-Uwe Pott 1, 2
Thomas Bertram 1
Andreas Eckart1
Christian Straubmeier 1
Christof Iserlohe 1
Wolfgang Voges 3
Günther Hasinger 3
1
I. Physikalisches Institut, Universität zu
Köln, Germany
2
ESO
3
Max-Planck-Institut für Extraterres-
trische Physik, Garching, Germany
The new VLT instrument SINFONI
gives us a view onto the circumnuclear
properties of AGN in unprecedented
detail, even beyond our local Universe.
As a science verification target, the
showcase object Mrk 609 demonstrates
impressively the necessity of adaptive
optics assisted integral-field spectros-
copy in order to distinguish between
Seyfert and starburst characteristics on
nuclear scales.
Fuelling of nuclear activity
The presence of the Seyfert phenomenon
is supposed to originate in the accre-
tion of matter onto a super-massive black
hole (SMBH) in the centre of a galaxy.
Nuclear activity is composed of nuclear
starbursts and Seyfert-like emission.
The fuel necessary for driving this activity
has to be transported from galactic
scales (~ 10 kpc) down to nuclear scales
of ~ 10 pc. We are still far from under-
standing the detailed processes that cre-
ate non-axisymmetric potentials leading
to the dissipation of angular momen-
tum. Such instabilities are needed for the
gas and stars to fall towards the nuclear
region. However, considerable theoretical
as well as observational effort has been
made to understand these processes
(e.g. Shlosman et al. 1990; Knapen 2005).
We can distinguish external and internal
triggers of the fuelling process:
External triggers are related to the envi-
ronment of galaxies and gravitational
interaction. Non-axisymmetry, which can
lead to loss of angular momentum, can
44 The Messenger 128 – June 2007
result from galaxy interactions. For Ultra
Luminous Infra-red Galaxies (ULIRGs),
which show the most extreme cases of
infrared nuclear activity, there is intriguing
evidence for a connection between
galaxy interaction and nuclear activity.
Internal triggers are based on instabil-
ities generated from within the host. For
example a two-step process has been
proposed that is able to sweep the inter-
stellar medium (ISM), via a stellar bar,
from large scales into a disk of several
hundred pc in radius. In the second step,
further instabilities (bar-within-bar) drive
the material close to the nucleus, until
viscous processes take over the angular
momentum transport.
While in quiescent galaxies, extensive
star formation appears to be related
to large-scale bars, the observational evi-
dence for non-axisymmetry-related
nuclear activity is not as clear for Seyfert
galaxies. A slight but significant increase
in the galactic-bar fraction of active gal-
axies has been found when compared to
non-active galaxies. This does not appear
to be the case for nuclear bars. Further-
more, there are a considerable number of
AGN that show no signs of the presence
of a bar, as well as of non-active galaxies
that do possess bars.
Mrk 609: a starburst/Seyfert composite
galaxy
Mrk 609 is classified as a starburst/
Seyfert composite galaxy. This class of
AGN appears to be best suited to study
the starburst-AGN connection, since
the AGN and starburst components pre-
sent themselves at the same level of
activity (Moran et al. 1996). Composite
galaxies can be characterised by optical
spectra which are dominated by star-
burst features, while the X-ray luminosity
and its variability are typical for Seyfert
galaxies. The former property is based on
the emission-line diagnostic diagrams
by Baldwin, Phillips and Terlevich (BPT
1981). Close inspection of the optical
spectra often reveals some weak Seyfert-
like features, e.g. [O iii] being significant-
ly broader than all other narrow lines, or a
weak broad Ha component. There is a
resemblance to narrow-line X-ray galax-
ies, which also show spectra of compos-
ite nature. Their soft X-ray spectra are
flat, but it is still not clear how this strong
and hard X-ray emission can be recon-
ciled with the weak optical Seyfert char-
acteristics. The faintness of these objects
in the X-ray, as well as in the optical
domain, has prevented them from being
studied in detail so far.
The SINFONI observations
Near-infrared imaging spectroscopy (cf.
Gillessen et al. 2006) has considerable
advantages over visible wavelength spec-
troscopy. Besides the much smaller dust
extinction, there are a number of NIR
diagnostic lines that probe the stellar and
non-stellar content in Mrk 609. Among
these are hydrogen recombination lines,
ro-vibrational transitions of H2, stellar fea-
tures like the CO(2–0) and CO(6–3) ab-
sorption band heads, as well as forbid-
den lines like [Fe ii] and [Si vi].
The data (Zuther et al. 2007) have been
acquired during the science verification
phase of SINFONI in August/October
2004. For the first time we have spatially
resolved the circumnuclear environment
on the scale of 270 pc in the J and H + K
bands (Figure 1). The morphology is com-
plex, and the continuum image reveals
a stellar bar-like structure superposed on
the point-like Seyfert nucleus (Figure 1a).
The distribution of hydrogen recombina-
tion emission (Paa) is clumpy and peaks
at the tip where the potential bar meets
the spiral arms and in regions along the
minor axis (Figure 1b). The presence of
nuclear broad Paa and [Si vi] are clear in-
dicators of the accretion of matter onto
a nuclear supermassive black hole. The
distribution of molecular hydrogen follows
the continuum shape, while that of [Fe ii]
is aligned with the minor axis of the
continuum and with the H-recombination
emission (Figure 1c and 1d).
The well established BPT emission line
diagnostics at visual wavelengths fail
for regions with considerable extinction.
Recently it was found that an analogous
NIR line diagnostic diagram, using [Fe ii]/
Pab and H2/Brγ line ratios, allows us to
distinguish between starbursts (photoion-
isation excitation), AGN (mixed excitation),
and LINERs (shock excitation). In this dia-
gram, the nucleus of Mrk 609 shows
signs of LINER activity (Figure 2). Shock-
driven excitation can be recognised by
its high [Fe ii]/Pab and H2/Brγ values. Our
integral-field data clearly resolve the nu-
clear and starburst activity in the central
kiloparsec. Extinction appears to play 3� × 3�
no crucial role in this region, since the
H-recombination line ratios are consistent K Paα
with unreddened case-B values. This is
furthermore supported by the strong Lya
emission in the ultraviolet.

The different regions characterised by

clumpy Paa emission (Figure 1b) clearly
follow a trend from a LINER-like value
N
at the nucleus to a starburst-like value in
the most distant North-eastern region
(data point 5 in Figure 2). The circumnu- [Fe II] E
clear regions (data points 2, 3, and 4) fall
in the domain of mixed excitation. In con- Figure 1: Hubble Space Telescope (HST) F606W-
band image of Mrk 609 (centre, from Malkan et al.
trast to our nuclear classification, the line
1998) overlayed with SINFONI maps of various
ratios extracted from the total field of spectral features: (a) K-band continuum map, (b) Paa
view resemble those of a typical Seyfert map, (c) [Fe ii] 1.275 µm, and (d) 1-0 S(1) molecular
galaxy. The same Seyfert characteris- hydrogen. The SINFONI field of view is indicated by
the white box in the HST image.
tics are obtained from the large aperture
Sloan Digital Sky Survey visual spectrum.
This demonstrates how the choice of
spatial scales to be studied can influence
the resulting classification. Our LINER
classification of Mrk 609, together with Shock
1.0
published data on variability of its non-
stellar NIR emission, might be explained 5
with the duty-cycle hypothesis, in which
0.5
short-lived accretion events occur period-
log ([Fe II]/Paβ)

ically and lead to the appearance of 1

Seyfert features in the high state and low- 2
0.0 total
ionisation (shock-driven) emission fea- LINER
tures in the low state. In order to ver-
4
ify this possible scenario, multi-epoch
– 0.5
SINFONI observations are needed. 3 AGN

The small spatial scales which SINFONI SB

–1.0
resolves on Mrk 609 allow for an investi- Photoionisation
gation of the transport of matter within
the central one kiloparsec. Obviously, –1.0 – 0.5 0.0 0.5 1.0 1.5
log (H 2 /Brγ)
Mrk 609 shows no signs of external trig-
gers (i.e. interaction with companion
galaxies). The presence of a nuclear bar- Figure 2: Line ratios of [Fe ii] 1.257 µm/Pab and Starbursts
like stellar structure, as well as the 1‑0 S(1) 2.121 µm/Brg. Activity types (starburst (SB), Seyfert 1
AGN and LINER) are indicated. Filled symbols re- Seyfert 2
clumpy morphology of the emission line present our SINFONI measurements of five distinct LINER
gas, fits into the picture of internal trig- regions of Mrk 609, as well as of the total field-of- Supernova
gers, feeding the AGN and circumnucle- view (red star). On account of missing J-band data,
ar star formation. However the chicken- region 5 is only accurately located along H2/Brg.
Open symbols correspond to literature values;
and-egg problem – whether the star orange diamonds represent starburst galaxies; red
for-mation activates the accretion onto left triangles Seyfert 1 galaxies; magenta headlong
the SMBH or AGN feed-back initiates triangles Seyfert 2s; blue circles LINERs; and grey
nearby starburst activity – remains open. asterisks supernovae. The diagram is interpreted as
displaying the transition from pure photoionisation
(lower left corner) to pure shock driven emission (up-
per right).

The Messenger 128 – June 2007 45

Astronomical Science
Outlook
The starburst/Seyfert composite nature
of Mrk 609 can be disentangled into dis-
tinct regions of star formation and strong
nuclear activity. Probably, this is true
for most composite galaxies. Whether
relationships between the occurrence of
dynamical instabilities, star formation,
and AGN activity are causally linked can
only be determined with AO-assisted
imaging spectroscopy (e.g. SINFONI) ob-
servations on a larger sample of this
class of objects. Moving beyond the
Gas
Gas velocity
46 The Messenger 128 – June 2007
backyard of our Milky Way towards larger
look-back times and stronger nuclear
activity, allows for a comparison with
local AGN on the same physical scales.
Then we are able to study scenarios such
as whether circumnuclear starburst
features generally accompany the cores
of galaxies with even stronger nuclear ac-
tivity, i.e. pure Seyferts and/or QSOs, or
if enhanced star formation is a distinct
phase in the evolution of AGN (e.g. Lípari
and Terlevich 2006). However, resolving
the central region of these objects on
the 100 pc scale is fundamental for these
studies.
1�
Stars E
Stellar velocity
References
Baldwin J. A., Phillips M. M. and Terlevich R. 1981,
PASP 93, 5
Gillessen S. et al. 2006, The Messenger 120, 26
Knapen J. H. 2005, Ap&SS 295, 85
Malkan M. A. et al. 1998, ApJS 117, 25
Moran E. C. et al. 1996, ApJS 106, 341
Rodríguez-Ardila A. et al. 2005, MNRAS 364, 1041
Shlosman I., Begelman M. C. and Frank J. 1990,
Nature 345, 679
Lípari S. L. and Terlevich R. J. 2006, MNRAS 368,
1001
Zuther J. et al. 2007, A&A 466, 451
The famous luminous infra red galaxy merger
Arp 220 is shown here in more results from VLT
SINFONI, this time used in tandem with the Laser
Guide Star System (LGS). The LGS provides real-
time adaptive optics correction using an artificial,
laser-fed star to correct the distortions of the at-
mosphere. The two nuclei of Arp 220, separated by
1.0 arsec, were resolved, the brighter to the west
(right) and the more diffuse to the east. The upper
images (where the brightness is colour-coded) show
the appearance in molecular gas (left) and stellar light
(right). The two lower images show the behaviour
of the velocities of the gas (left) and the stars (right);
the colour-coded images here depict matter mov-
ing towards the observer (in blue) and moving away
(red). It is apparent that the stars and the gas move
differently: the stars in two counter-rotating discs
and the gas in a larger-scale disc. The image is taken
from ESO PR 27/07, which provides more details.
Astronomical Science
GHostS – Gamma-Ray Burst Host Studies
Sandra Savaglio 1
Tamás Budavári 2
Karl Glazebrook 3
Damien Le Borgne 4
Emeric Le Floc’h 5
Hsiao-Wen Chen 6
Jochen Greiner 1
Aybuk Küpcü Yoldaş1
1
 ax-Planck Institute for Extraterrestrial
M Today GRBs are primarily discovered by
Physics, Garching, Germany
2
Johns Hopkins University, Baltimore,
USA
3
Swinburne University, Melbourne,
Australia
4
CEA-Saclay, Gif-sur-Yvette Cedex,
France
5
Institute for Astronomy, Honolulu,
Hawaii
6
University of Chicago, USA
GHostS is the largest public data-base
on gamma-ray burst (GRB) host galax-
ies and is accessible at the URL http://
www.grbhosts.org. Started in 2005, it
currently contains photometric and
spectroscopic information on 39 GRB
hosts, almost 2/5 of the total number
of GRBs with measured redshift. It will
continue to grow, together with the un-
stoppable data flow from the obser-
vatories all over the world, every time
a new event is discovered. Among
other features, GHostS uses the Virtual
Observatory resources.
The Gamma-Ray Burst phenomenon
Gamma-ray bursts (GRBs) are the most
energetic events in the Universe, and
also among the fastest. They are associ-
ated with the death of massive stars (core
collapse supernovae) or the merging of
compact objects, such as neutron stars
and black holes. During the explosive
phase, they emit most of their energy as
Number
a collimated flux of gamma-ray photons
from neutrino-antineutrino annihilations,
and as gravitational waves. The gamma
emission lasts from a few milliseconds to
a few minutes.
The first GRB was discovered in 1967, by
the US military satellite Vela, but it took
six years, until 1973, for the first paper to
appear in a scientific journal. The first red-
shift was measured in 1997, when GRBs
were finally confirmed to be cosmologi-
cal sources. Today there are 108 GRBs
with measured redshift, the highest being
GRB 050904, at z = 6.3. The typical
energy emitted by a GRB, in a couple of
minutes, is 10 51 ergs, equivalent to the
energy emitted by the sun in 10 billion
years.
the satellite Swift (http://swift.gsfc.nasa.
gov), the dedicated NASA mission which
in two years of operation helped to double
the number of measured redshifts. With
the growing data collected from space
and ground telescopes, and the advent
of Swift, our group decided to create a
database, available to scientists, which
collects observational results on the gal-
axies hosting GRB events. We called it
GHostS, or GRB Host Studies. Our focus
is to explore and unveil the nature of gal-
axies in which GRBs occur. Our goal is to
answer fundamental questions, such as:
are these galaxies different from the gen-
eral galaxy population; are they special in
any way; can we use them to understand
galaxy formation under extreme condi-
tions?
Today the GHostS database includes
results for 39 GRB host galaxies. This is
a large number, considering that only
108 GRBs have known redshift (see Fig-
ure 1).
12
10
8
6
4
2
0 GRBs (empty histogram) and the
0 1 2 3 4 5
Redshift
The GRB host galaxy population
There is no doubt that studies of GRB
host galaxies are entering into a realm of
galaxy formation that was hardly known
before. The reason is that GRB hosts
are generally faint and detected at high
redshifts. Similar galaxies are very hard
to find using conventional techniques,
because they would require very long
integration times, even for the largest and
most efficient telescopes.
GRB events offer a shortcut to the quest
for faint and distant galaxies. They are
detected as short and energetic events.
The localisation (and for a fraction of
them the redshift) is measured from the
bright X-ray and optical afterglow. We
know that long-duration GRBs occur in
regions of star formation, therefore, they
are associated with galaxies. Dedicat-
ed programmes can observationally com-
plete the investigation.
In our GHostS search, we want to explore
many of the most interesting galaxy pa-
rameters, such as metallicity, star-forma-
tion rate, stellar mass, age of the stellar
population and dust extinction. Each of
these parameters are very difficult to de-
rive. One difficulty is the faintness of the
typical GRB host. Moreover, the data ob-
tained by the community are often not
homogeneous. To complicate the picture,
the tools often used to derive the physical
quantities are affected by systematic un-
certainties which are sometimes greater
than the relations being derived.
Figure 1: Redshift distribution of all
6 7 subsample of GRBs included in the
GHostS database (blue histogram).
The Messenger 128 – June 2007 47
Astronomical Science Savaglio S. et al., GHostS – Gamma-Ray Burst Host Studies

The typical GRB host is a star-forming,
low-mass and low-metallicity galaxy, de-
tected at redshift below z = 2. The mean
stellar mass we derived in our sample
is similar to the stellar mass of the Large
Magellanic Cloud (Figure 2). Observation-
al limitations prevent us from fully explor-
Number
9 Figure 2: Histogram of the stellar mass
of GRB host galaxies in the redshift
8 interval 0 < z < 6.3, from the GHostS
sample. The mean stellar mass is
7 of the order of the stellar mass of the
Large Magellanic Cloud. This is ten
times smaller than the stellar-mass
6 limit reached by the deepest high-
redshift galaxy surveys, performed by

ing the same parameters in GRBs at 5 todays’ largest telescopes.

larger redshift, but in exceptional cases.
4
The detection of the cold interstellar me-
dium in high-redshift hosts indicates that 3
there the population is different, with
larger stellar masses, higher star-forma- 2
tion rates and higher metallicities (Berger
1
et al. 2005, Fynbo et al. 2006, Savaglio
2006). It is still unclear if this means a 0
different population, or different observa- 7 8 9 10 11 12
tional biases. Such issues will be faced log M �(M � )
and fully exploited by the future genera-
tion of telescopes and instruments.
Figure 3: Example of the
GHostS database web
In general, we know that GRB hosts are
pages. Each GRB entry
very peculiar galaxies. We still do not includes basic informa-
know whether this is because our tradi- tion on the event, togeth-
tional observational technique cannot er with access to rele-
vant papers available in
reach the extreme limits attained by GRB
the literature. It reports
detections, or because GRB hosts are in- multi-band photometry,
trinsically different. The main limitation is fluxes of emission lines
the small number of galaxies discovered and images of the field.
so far. This number is still below 100, while
the number of galaxies in today’s sur-
veys exceed 10 5.

The GHostS database

GHostS includes GRB hosts for which

photometric and spectroscopic informa-
tion is available from the literature (Fig-
ure 3). In total 39 galaxies, out of a total
of 108 GRBs whose redshift is known,
have been gathered from 81 different pa-
pers and sources. 7 10 –17 Figure 4: Very Large
Telescope spectrum of
the galaxy hosting
One of the spectra in the GHostS data- 6 10 –17 GRB 990712 at redshift
base is shown in Figure 4. The galaxy z = 0.433 (Küpcü Yoldaş
hosting GRB 990712 at redshift z = 0.433 5 10 –17 et al. 2006).
was observed at the Very Large Telescope
Flux (erg/cm 2 /s/Å)

(Küpcü Yoldaş, Greiner, Perna, 2006). It 4 10 –17

has very strong emission lines, originating
[O III]
[O III]

in regions of the galaxy where star forma- 3 10 –17

tion is very active. We derived a stellar
mass of the galaxy of a few times 109 MA,
[O II]

2 10 –17
Hβ

similar to the stellar mass of the Large

[Ne III]

Magellanic Cloud, but it forms stars

Hγ

1 10 –17
at a rate that is ten times higher, about
6 MA yr –1. This gives a specific star for-
0
mation (the SFR per unit stellar mass) of 5 000 5 500 6 000 6 500 7000 7500
100 times that in the Milky Way. Similar Wavelength (Å)

48 The Messenger 128 – June 2007

galaxies hosting GRBs are detected up to
a distance when the Universe was less
than 1 Gyr old. They are very hard to find
using normal tools of investigation. Most
surveys, probing the high-redshift Uni-
verse today, can reach, in terms of stellar
mass, galaxies similar to or larger than
the Milky Way.
It is generally not easy to collect results
published by the GRB community, be-
cause the number of papers available for
each event can be large. One extreme
case was GRB 060218, which occurred
in February 2006, at redshift z = 0.0335,
associated with a supernova, detected
three days after the alert (Masetti et al.
2006). In about a year and a half, more
than 20 articles dedicated to this event
have been published (or are in the proc-
ess of being published) in refereed jour-
nals, four of which appeared in Nature.
This proliferation of material is one of
the main motivations for the existence of
GHostS. Observational results are
searched, collected, homogenised and
made easily accessible for the whole
community. The database is constantly
growing in terms of total number of
objects, and in terms of tools offered. We
plan to eventually include all host galax-
ies discovered in the past and in the fu-
ture for the years of operation of Swift,
and it will likely contain a final sample of a
few hundred galaxies.
The Virtual Observatory
The Virtual Observatory (VO) is one of the
features offered in our database. The
VO is one of the newest and most prom-
ising tools introduced in the astronomy
world, allowing scientists to easily access
data from multiple astronomical obser-
vatories, both ground- and space-based
facilities, through one portal (see for in-
stance http://www.us-vo.org/). Its com-
ponents have been added to the GHostS
web portal (Figure 5). Using information
about the GRBs and their hosts stored
inside our SQL database, one can auto-
matically enhance the view of the cata-
logued hosts by using various VO re-
sources. The details pages currently show
images of the relevant pieces of the sky
as seen by the Sloan Digital Sky Survey
(SDSS) and the legacy Digitized Sky Sur-
1� N
E W
S
vey (POSS, SERC). Other types of auto-
mated searches are also in the works,
e.g., for finding spectra at the given posi-
tion. These VO additions, the efficient
database, and the presentation of the
scientific data, make the website not only
an invaluable research tool for GRB host
studies, but also very enjoyable to navi-
gate. GHostS will feature in the future ad-
vanced search capabilities for the com-
munity, fine-tuned for astronomical data
and will eventually dynamically search VO
resources for relevant observations, such
as images and spectra.
The future
The future of GRB science fully depends
on the life of GRB hunting machines,
i.e. gamma-ray detectors. Today it almost
exclusively relies on the existence of
Swift, which has been funded for an ad-
ditional four years.
Among the many unexplored territories,
we mention the discovery of the very
high redshift GRBs, at z > 6. We know
that there are collapsed objects at those
times. We do not know under which
physical conditions these objects formed:
are they stars, galaxies or massive black
holes? At z > 6, QSO searching has failed
in providing a satisfactory picture, and
galaxies are incredibly hard to find. GRBs
on the other hand have shown some
very promising results. In the first two
years of operations, Swift has discovered
8 GRBs at redshift larger than 4, and 4
at redshift larger than 5. This is 7 % and
3.5 % of the total GRB population with
measured redshift, and probably the num-
Figure 5: One example of the tools offered
by GHostS. The images show the field of
GRB 050509b, at redshift z = 0.2248 (Gehrels et al.
2005), obtained from the Sloan Digital Sky Sur-
vey (left) and Digitalized Sky Survey (right) archives,
through the Virtual Observatory.
bers would be larger if we were able to
observe all GRB afterglows early on. In
fact, 2/3, of the Swift GRBs have no mea-
sured redshift. The fractions of z > 4 and
z > 5 known QSOs are 0.8 %. and 0.03 %
of the total, with only 17 QSOs known
with redshift larger than 5.
Particularly impressive was GRB 050904,
the record holder at z = 6.3 (Kawai et al.
2005). This GRB was observed with the
Japanese telescope Subaru more than
three days after the Swift alert, when it
had already faded 11 magnitudes from its
initial brightness. Fortunately at high red-
shifts the brightness of a fading object
is helped by time dilation, the effect de-
scribed in Einstein’s theory of General
Relativity. If this GRB had occurred in the
local Universe, it would have faded at
a rate more than seven times faster. The
spectrum of this spectacular event is
shown in Figure 6 (from Totani et al. 2006).
The host galaxy is barely detected by the
Hubble Space Telescope and Spitzer
Space Telescope (Berger et al. 2006). We
estimate a stellar mass which is half
that of the Large Magellanic Cloud, and a
rate of star formation ten times higher. Al-
though this event happened less than
900 million years after the Big Bang, the
absorption lines detected in its spectrum
indicate a surprisingly high pollution of
heavy elements in the interstellar medium
in the host galaxy, of the order of 1/10
that of the solar vicinity (Totani et al. 2006;
The Messenger 128 – June 2007 49
Astronomical Science

Figure 6: Subaru spectrum of the most distant GRB

ever discovered, GRB 050904 at z = 6.3 (Totani et al.
2006). It was observed more than three days after
the alert given by the gamma-ray burst finder Swift,
when it had already faded 11 magnitudes from its
initial brightness.

Savaglio 2006). This is hardly explained

by most theories of chemical enrich-
ment in the primordial Universe. Galaxies 1.5
like these are very likely faint and dormant
today, because they would have con-
sumed their gas reservoir for star forma- 1
Fλ(10 –18 erg cm –2 s –1 Å –1)
tion in less than one Gyr, i.e. by redshift
z = 3.
0.5
If GRB 050904 were observed two hours GRB 050904
after the Swift alert, its spectrum would 3.4 days
0
have been much more spectacular, with
a signal-to-noise ratio ten times better
than that of the Subaru spectrum. Noth- – 0.5 (C IV) S II OI C II
ing like this has ever been achieved in the Lyα + Si II
1σ errors
remote Universe with normal galaxies
or QSOs. It is only a matter of time that, –1
8 600 8 800 9 000 9 200 9 400 9 600 9 800 10 000
sooner or later, Swift will trigger similar
Observed wavelength (Å)
events at higher redshift. Then telescopes
from the ground will be ready to catch
the fading sources and deliver a unique
data set to the scientific community, for References Kawai N. et al. 2005, GRB Coordinates Network
3937, 1
new exciting discoveries. We will be there
Berger E. et al. 2005, ApJ 642, 979 Küpcü Yoldaş A., Greiner J. and Perna R. 2006,
to continue our service to the community Berger E. et al. 2006, ArXiv Astrophysics e-prints, A&A 457, 115
with GHostS, in the attempt to fully char- arXiv: astro-ph/0603689 Masetti N. et al. 2006, GRB Coordinates Network
acterise those galaxies parenting one of Fynbo J. P. U. et al. 2006, A&A 451, L47 4803, 1
Gehrels N. et al. 2005, Nature 437, 851 Savaglio S. 2006, New Journal of Physics 8, 195
the most extraordinary objects in the Uni-
Totani T. et al. 2006, PASJ 58, 485
verse.

The Rapid-Eye Mount

(REM) 0.6-m telescope
at the ESO La Silla Ob-
servatory, which is dedi-
cated to the follow-up
of gamma-ray bursts,
is shown in operation.
ESO PR 26/07 provides
details of the telescope
and some science re-
sults.

50 The Messenger 128 – June 2007

Astronomical Science
The Puzzle of the Lya Galaxies: New Results from the VLT
Christian Tapken 1
Immo Appenzeller 2
Armin Gabasch 3, 4, 5
Jochen Heidt 2
Ulrich Hopp 3, 4
Ralf Bender 3, 4
Stefan Noll 3, 4
Stella Seitz 3, 4
Sabine Richling 6
1
Max-Planck-Institut für Astronomie,
Heidelberg, Germany
2
Landessternwarte Heidelberg-König-
stuhl, Heidelberg, Germany
3
Universitäts-Sternwarte München,
Germany
4
Max-Planck-Institut für Extraterres-
trische Physik, Garching, Germany
5
ESO
6
Institut d’Astrophysique de Paris, France
Observations of high-redshift galaxies
show that at early cosmic epochs the
cosmic UV radiation field appears to be
dominated by small galaxies with strong
Ly-alpha emission. Although usually
small and of relatively low luminosity,
these galaxies are easily identified from
their line emission. Observations with
the VLT resulted in significant progress
in the understanding of the nature of
these distant galaxies and of their role
in the early Universe.
The pioneers: R. B. Partridge and
P. J. E. Peebles
The history of the Lya galaxies began in
1967 with a pioneering paper by R. B.
Partridge and P. J. E. Peebles. In this arti-
cle, published in the Astrophysical Jour-
nal, the two Princeton University astro-
physicists discussed the formation of the
first galaxies in the Universe. They con-
cluded that the first galaxies probably
began their lives with strong initial bursts
of star formation at cosmic epochs cor-
responding to redshifts between 10 and
30. Because of the presence of many
massive, hot, and luminous stars in these
starbursts, Partridge and Peebles pre-
dicted very high ultra-violet (UV) luminosi-
ties of the newly formed galaxies, and
they estimated that the light emitted by
such objects during the first few hundred
million years should actually be observ-
able. Because of the large distance, the
light emitted by the young galaxies as far-
UV radiation is redshifted to the red and
near-infrared spectral range. Hence, the
detection of such objects requires obser-
vations at these wavelengths.
As the most promising way to find these
objects, Partridge and Peebles proposed
to search for their Lyman-a (Lya) line
emission. This spectral line, due to transi-
tions from the first excited energy level
to the ground state of hydrogen atoms, is
important in many astrophysical proc-
esses. In normal galaxies Lya emission is
produced during the recombination of
interstellar hydrogen gas, which has been
ionised by hot stars. Simple estimates
show that up to about 2/3 of the Lyman
continuum photons (and up to about 6 %
of the total luminosity) of hot stars can
be converted into Lya photons. This also
means that, in principle, Lya emission
equivalent widths of the order 100 to
200 Å can be expected in the spectra of
such galaxies.
Of course 6 % is still only a minor fraction
of the total emitted luminosity. However,
as noted by Partridge and Peebles, in the
small wavelength range covered by the
Lya line, the expected spectral flux densi-
ty is much higher than in the adjacent
continuum. Therefore, detecting the line
emission against the strong sky back-
ground in the red spectral range ap-
peared much easier than looking for the
continuum emission.
In 1967 sensitive astronomical detectors
were limited to wavelengths ≤ 1 µm. Thus,
observations of the redshifted Lya line
appeared feasible only up to redshifts
where the observed wavelength of Lya
does not exceed 1mm, which means red-
shifts z = ∆λ/λ ≤ 7. In view of this limita-
tion, Partridge and Peebles calculated
the emitted luminosity and the expected
observed Lya flux emitted at an epoch
corresponding to z ≈ 7.
To appreciate the courage and foresight
of taking up this topic in 1966, when the
paper was submitted, we have to recall
that the Cosmic Microwave Background,
confirming the present cosmological
concepts, had been discovered just one
year before, that the most distant galaxy
known at that time had a redshift of
about 0.5, that neither galaxy formation
nor star formation was well understood,
that the most advanced astronomical
detectors were image tubes followed by
photographic plates, and that there ex-
isted only two astronomical telescopes
with apertures exceeding 2.5 m. Since in
1966 getting time at the largest existing
telescopes was probably as difficult as
today, Partridge and Peebles assumed
for their feasibility estimates a more mod-
est (and more typical) 90-cm telescope,
equipped with an image tube with an
S1 cathode. With these assumptions
­Partridge and Peebles predicted that the
Lya emission of very young galaxies
at z ≈ 7 should be detectable with an
exposure time as short as five minutes(!).
The search for the Lya emitting galaxies
Prompted by the 1967 paper, many dif-
ferent groups started searching for red-
shifted Lya (and UV continuum) emission
of high-redshift galaxies. However, al-
though a large amount of observing time
was invested, and although the surveys
soon reached much fainter magnitudes
than those predicted by Partridge and
Peebles, for many years no redshifted
Lya emitting galaxy was found. Some
Lya emission was detected from distant
radio galaxies and from some galaxies
associated with distant QSOs. However,
it was not clear whether starbursts or
non-thermal ionising sources were re-
sponsible for the Lya emission from these
objects.
Starting about 1995, many galaxies with
redshifts of z ≈ 3 were discovered using
the Lyman-break technique. Practically all
the distant galaxies found with this tech-
nique showed clear spectroscopic sig-
natures of strong starbursts. However, in
most cases the Lya line occurred ei-
ther in absorption or as a relatively weak
emission feature. Similar results were
found for samples of high-redshift gal-
axies selected on the basis of photomet-
ric redshifts (see Figure 1).
More detailed models of starburst gal-
axies soon provided a plausible explana-
tion for the weakness of the Lya emis-
sion of the young z ≈ 3 galaxies. Partridge
and Peebles were certainly correct,
noting that in young starbursts copious
The Messenger 128 – June 2007 51
Astronomical Science
numbers of Lya photons are produced.
However, because of the high absorption
cross-section of hydrogen atoms for Lya
photons, even a small amount of neutral
hydrogen can make a galaxy completely
opaque for Lya.
Normally the absorption of Lya by an in-
terstellar hydrogen atom will be followed
by a re-emission at the same frequency.
Thus, hydrogen atoms essentially scat-
ter the Lya photons, which in principle
can still escape after some random walk
through the neutral hydrogen layer. How-
ever, the resonance scattering greatly
increases the effective light path of these
photons. Therefore, even a small amount
of truly absorbing material embedded in
the scattering layer will result in the even-
tual absorption of the Lya photons. A
most efficient UV absorber is interstellar
dust, which is normally abundant in star-
forming regions. Thus, the combination of
resonance scattering by neutral hydrogen
and dust absorption provides a highly
plausible explanation for the absence
of strong Lya emission in most known
young galaxies. It also seemed to explain
why more than two decades of searches
for high-redshift Lya galaxies were not
successful.
Soon after the above explanation for the
absence of strong Lya emission in dis-
tant starburst galaxies, and the reason for
the futility of the earlier searches had
been more or less accepted, the history
of the Lya galaxies took an unexpected
turn in the mid-1990’s. The first Lya gal-
axies were detected on the basis of their
strong emission line. In an ApJ letter
of 1998 Ester Hu, Lennox Cowie, and
Richard McMahon reported the discovery
of the elusive high-redshift Lya galaxies
during observations with the new 10-m
Keck II telescope. The objects discovered
by these authors had redshifts of z ≈ 3.4
and z ≈ 4.5 and Lya emission equivalent
widths > 100 Å. However, their luminos-
ities were only about one hundredth of
that predicted by Partridge and Peebles.
This obviously explains why it took a tele-
scope of 10 m (instead of 1 m) aperture to
find them.
From the large Lya equivalent widths, it
was clear that the faintness of these
objects was not due to dust absorption.
Obviously, these galaxies were intrinsi-
52 The Messenger 128 – June 2007
Tapken C. et al., The Puzzle of the Lya Galaxies: New Results from the VLT
25
20
15
Frequency
10
5
0
0 100
Lyα emission equivalent widths (Å)
cally faint and small (see Figure 2). On the
other hand, the relatively high space den-
sity of the distant Lya galaxies derived by
Hu et al. indicated that the Lya galaxies
provided a major contribution to the star
formation and to the cosmic radiation
field at the corresponding epochs. Never-
theless a reliable assessment of the ef-
fects and the importance of the Lya gal-
axies required more information on their
detailed properties and their density as a
function of redshift. Therefore, the first
results immediately started a new wave of
searches for high redshift Lya galaxies.
In addition, new studies of the physics of
these objects were triggered. Some of
this new work was carried out at ESO us-
ing the VLT. In the following we describe
new results on the nature and the space
density of these distant galaxies which
resulted from these VLT observations.
The mystery of the strong line emission
As described above, for many years as-
tronomers were puzzled by the fact that
no Lya galaxies could be observed. That
they were finally discovered in the mid-
1990’s resulted in a new big mystery. The
question now was, why these galaxies
could emit such a strong Lya flux in spite
of the expected quenching of the Lya
photons by the combination of resonance
scattering and dust absorption. As a pos-
sible answer to this question it was sug-
gested in the literature that the Lya galax-
ies were too young to have formed heavy
chemical elements, which are needed
to form dust. This explanation appeared
plausible, since the first galaxies are ex-
pected to form from primordial matter
produced by the Big Bang, which (apart
from tiny amounts of deuterium and
Figure 1: Distribution of the Lya
emission equivalent widths in samples
of high-redshift galaxies selected
using broadband photometry. The
histogram is based on the FDF
spectroscopic survey (Noll et al. 2004),
selected on the basis of photometric
redshifts. The blue line shows the
distribution derived by Shapley et al.
(2003) for Lyman-break-selected
galaxies.
200 300
lithium) is composed of hydrogen and he-
lium. Under astrophysical conditions
neither hydrogen nor helium can form
solids. Thus, dust particles can form only
after heavier chemical elements have
been produced by the nuclear processes
in stars. A very important first result of
the VLT observations was high-quality
FORS spectra of Lya galaxies showing
conspicuous lines of heavy elements (Fig-
ure 3). In particular, these spectra con-
tained lines of carbon and silicon, which
are among the main constituents of inter-
stellar dust particles.
If dust is present in young galaxies, Lya
photons can still escape, if the dust and
(or) the neutral hydrogen are distributed
inhomogeneously and if, by chance, no
such material is present along the line of
sight between us and the starburst. That
the amount of dust and cool material
along the line of sight plays a role for the
observed Lya flux is supported by the
finding that the Lya emission strength of
high-redshift galaxies decreases with
increasing dust reddening and that the
Lya emission is lower in galaxies with
strong interstellar absorption lines, typical
for cool interstellar gas (Shapley et al.
2003, Noll et al. 2004). However, for a
given reddening, the Lya emission
strength tends to vary considerably and
there exist Lya galaxies with high Lya
equivalent widths and a modest amount
of reddening. Thus, missing dust and/or
missing neutral hydrogen along the line
of sight cannot fully explain the observed
strong line emission.
Another mechanism, which can avoid the
quenching of the Lya emission, is the
presence of sufficiently large velocity
gradients in the neutral gas. If a hydrogen

FDF-4691
5 kpc
10
8 profiles and the expansion of a neutral
Ly–lim
Lyβ
Lyγ
Flux (10 –21 Wm –2 Å –1)
6
4 the emitting region.
2
0
4 000
atom is moving relative to the volume
emitting the Lya line with a velocity larger
than the line widths, the atom cannot
absorb or scatter the Lya photons. In this
case the Lya photons can penetrate the
neutral hydrogen layer and escape. Such
velocity differences are not unexpected
since the velocity of sound is much
higher in the hot Lya emitting gas than in
the cool neutral hydrogen layers. Thus,
assuming turbulent media with subsonic
turbulence, the Lya lines emitted by
the hot ionised gas are expected to have
profiles which are significantly broader
than the velocity dispersion of the scat-
tering layers. However, because of the
very high absorption cross-section of
neutral hydrogen, even the outer wings of
the absorption profiles can prevent the
escape of the Lya photons. On the other
hand, it can be shown that the probability
for an escape of the Lya photons can
increase strongly if large-scale velocity
fields are present in the scattering layers.
Mechanisms producing such large-scale
velocity fields can be mass infall, or mass
outflows from the galaxies powered by
stellar winds or radiation pressure from
the central stars. Since such velocity
fields also affect the profiles of the Lya
lines, line profile observations provide a
critical test of this hypothesis.
FDF-3389
5 kpc 5 kpc
Figure 3: FORS spectrum of the Lya
galaxy FDF-4691 (from Tapken et al.
FDF-4691
2004).
Lyα
Si IV
C IV
NV
C II
5 000 6 000 7000
Wavelength (Å)
We have used the VPH grisms available
on the FORS instruments to obtain
medium-resolution spectra of a sample of
Lya emitting high-redshift galaxies in
the FORS Deep Field (Appenzeller et al.
2004). The observed Lya profiles were
compared to model profiles, which were
computed using the radiative transfer
code of Meinköhn and Richling (2002),
which is particularly well suited for mod-
elling resonance scattering in moving
media. Examples of the observed profiles
and model fits are presented in Figure 4.
For these computations (described in
detail in Tapken et al. 2007) we assumed
a central volume of turbulent, ionised
hydrogen, emitting the Lya radiation. This
region was assumed to be surrounded by
a shell of dusty neutral hydrogen.
The calculations showed that the ob-
served profiles could be reproduced rea-
sonably well with this model, if a modest
(10–200 km s –1) expansion velocity of
the neutral hydrogen shell was assumed.
From the observed line wings (which are
not much affected by the neutral hydro-
gen shell) high velocities (> 600 km s –1) of
the ionised gas could be derived. These
high velocities can be explained by super-
sonic turbulence caused by supernova
shocks and strong stellar winds in the
Figure 2: Examples of HST/ACS
FDF-5812 F814W images of three Lya galaxies.
compact, strong starbursts. The combi-
nation of the relatively broad intrinsic Lya
hydrogen shell readily explains the fact
that a major fraction of the Lya photons
can escape, in spite of the presence of
a dusty neutral hydrogen shell in front of
From our results, we conclude that, al-
though a relatively small amount of
neutral matter along the line of sight cer-
tainly plays a role, the high intrinsic
velocity dispersion of the emitting ionised
gas and large-scale outflows of the
neutral hydrogen are the main reasons for
the escape of a large fraction of the Lya
photons from the observed galaxies. The
small continuum flux (relative to other
high-redshift galaxies) indicates a low
total mass of the hot stars of Lya galax-
ies. If the mass of the hot stars is a mea-
sure of the total mass of these objects,
Lya galaxies are probably of lower (and
more normal) mass than other known
high-redshift galaxies. Therefore, they
probably have lower gravitational poten-
tials, which may explain the observed
strong outflows.
The space density of Lya galaxies and
their role in cosmic evolution
As noted above, most known Lya galax-
ies are small and have intrinsically faint
UV continua. However, already Hu et al.
(1998) found for the redshift range 3.4 <
z < 4.5 a total star-formation density due
to the Lya galaxies which is comparable
to that due to all other known high-red-
shift galaxies. Since Lya galaxies tend to
be less affected by reddening than other
high-redshift galaxies, they are expected
to dominate the UV radiation field at these
redshifts. Moreover, studies of high-red-
shift galaxy samples based on photomet-
The Messenger 128 – June 2007 53
Astronomical Science
Figure 4: Examples of observed (black) and
computed (red) Lya emission line profiles of
Lya emitting galaxies.
FDF-4691
15
10
Flux
5
0 0
– 2 000 –1000 0 1000
Velocity (km s –1)
ric redshifts (and, therefore, relatively free
of selection effects) show that the fraction
of Lya emitting galaxies is increasing with
redshift (Noll et al. 2004). According to
Shimasaku et al. (2006) at z = 6 about
80 % of all high-redshift galaxies are Lya
galaxies with intrinsic Lya emission equiv-
alent widths > 100 Å. Hence, at very
high redshifts these objects are expected
to strongly dominate the cosmic UV radi-
ation field and the cosmic ionisation.
For a more quantitative assessment of
the role of the Lya galaxies in the early
Universe obviously a reliable knowledge
of their space density as a function of
luminosity and redshift is needed. There-
fore, during the past years, various differ-
ent groups have invested much work to
derive this so-called ‘luminosity function’
of the Lya galaxies at different redshifts.
There were basically two types of such
programmes: firstly, large-area surveys
were used to improve the number statis-
tics of these objects; secondly, very deep
observations in smaller fields were used
to reach fainter Lya galaxies and to derive
the faint part of the luminosity function for
these objects. This faint part is important,
since, within the observational limits, faint
galaxies are always more numerous than
the bright ones.
Particularly successful among the large-
area surveys were studies carried out
with the wide-field SUPRIME camera of
the Japanese national telescope Subaru
(see, e.g., Shimasaku et al. 2006). These
surveys resulted in important informa-
tion on the space density of luminous Lya
galaxies and on the cosmic variance of
this quantity. Our contribution to the topic
54 The Messenger 128 – June 2007
Tapken C. et al., The Puzzle of the Lya Galaxies: New Results from the VLT
20
FDF-5215
15
Flux
10
5 1
2 000 –1000 0 1000
Velocity (km s –1)
was an extension of the luminosity func-
tion to lower luminosities, which was pos-
sible as a result of the superior sensitivity
of the FORS2 instrument and a spe-
cially developed set of narrowband filters
(Tapken et al. 2006).
Like all current searches for Lya galaxies,
the survey carried out with the VLT used
sky images obtained through a combina-
tion of broadband and narrowband fil-
ters. Normal galaxies, where the light is
dominated by the stellar continua, tend
to be visible with a similar brightness in
many different filter bands. Emission-line
objects are characterised by an excess
emission in one or, if the emission line co-
incides with the overlap region of two
filters, in at most two adjacent filter bands
(see Figure 5). Filter photometry allows a
reliable detection of galaxies which have
emission lines in their spectra. More dif-
ficult and more complex is the unambigu-
ous identification of the observed emis-
sion lines as Lya. For this purpose one
needs reliable photometric redshifts and
low-resolution spectra of the candidate
galaxies, which rule out other identifica-
tions of the observed lines.
In order to reach an optimal signal-to-
noise ratio, the passbands of the narrow-
band filters used for searches of Lya
galaxies are designed to coincide with
wavelength regions of particularly low
night sky background. Since most of the
night sky background in the red and
near-infrared is due to airglow produced
by OH molecules in the high atmosphere,
the spectral regions of low sky back-
ground are the gaps or ‘windows’ in the
OH line spectrum. One of these ‘OH win-
FDF-7539
4
3
Flux
2
0
–1000 0 1000 2 000
Velocity (km s –1)
dows’ covers about 20 nm wavelength
interval near l = 815 nm (corresponding
to a Lya redshift of z ≈ 5.7). This win-
dow is particularly important since, at this
wavelength, photometric and spectro-
scopic follow-up observations are still rel-
atively easy. Therefore, this window was
also used for our VLT observations. To
reach an optimal sensitivity, the wave-
length interval of the 815-nm OH window
was covered by a set of three filters (Fig-
ure 6). In this way it is possible not only to
avoid the strong OH lines outside the
window, but also to partially suppress the
weaker lines still present inside the
815-nm OH gap. Because of the resulting
lower background, the VLT observations
allowed us to reach significantly lower
Lya luminosities than had been possible
before.
Figure 7 shows the results together with
luminosity function data from another
recent survey. The new data confirm the
high space density of faint Lya galaxies
at high redshift. In particular up to z ≈ 6
the space density seems not to decrease
with redshift. On the other hand, the ob-
served luminosity range and the number
of observed objects are still too small, to
constrain the luminosity function well.
Although much progress has been made,
obviously more work is needed to derive
the radiation field produced by the Lya
galaxies with an adequate accuracy.
Future work and outlook
Several ambitious large-area searches for
Lya galaxies are underway which will
further improve the statistics of these ob-
 Figure 5: Bessell-R and broadband images and
narrowband images of four galaxies in the FORS
Deep Field. The upper two panels show z ≈ 5.7
Lya galaxies. The lower two panels show, for com-
parison, corresponding data for a normal z ≈ 3.2
starburst galaxy and for a z ≈ 1.2 [O ii] emitter.

jects. Of particular interest will be the R I 8100_70 8150_50 8150_60

results of the DAzLE survey, which, using
infrared imaging, aims at finding Lya gal-
axies at redshifts z > 7. Hence, during the FDFLAE-12
next few years we will certainly see much
more and better statistical data on the
bright part of the luminosity function of
the Lya galaxies. Less clear is the out-
look for more information on the impor-
tant faint part of the luminosity function of FDFLAE-13
these objects. Reaching significantly
fainter galaxies with normal deep-field
observations would require very long ex-
posures times, which appear not realis-
tic for such programmes. More promising
may be the ‘gravitational telescope’ tech- FDF-1999
nique, i.e., making use of the flux amplifi-
cation of distant objects by strong gravi-
tational lensing. Because of their relatively
high surface density, Lya galaxies are well
suited for such programmes. Estimates
have shown that, with presently available FDF-5115
instruments, observations of lensed Lya
galaxies in the field of a single galaxy
cluster with a high lensing strength (such
as the ROSAT source RX J1347-1145)
could provide a significant sample of Lya
galaxies with intrinsic luminosities one to 4 Figure 6: Spectrum of the OH airglow
near the 815 nm window. Overplotted
several magnitudes fainter than those
are the transmission curves of the
known at present. These data could be special FORS2 narrowband filter set
obtained with a rather modest amount of 3 for detecting Lya galaxies at redshifts
observing time. Thus, progress concern- ≈ 5.7.
ing the luminosity functions also seems to
be possible, provided observing time
Flux

2
for such programmes can be obtained.

1
Acknowledgements

Our research has been supported by the German

Science Foundation DFG (SFB 439). We would like 0
7900 8 000 8100 8 200 8 300 8 400
to thank the staff of the Paranal Observatory for Wavelength (Å)
carrying out the service mode observations. We also
thank Dörte Mehlert and Erik Meinköhn for discus-
sions and valuable comments.

–2 Figure 7: The luminosity function of

References Lya galaxies at z = 5.7.
Tapken et al. 2006

Appenzeller I. et al. 2000, The Messenger 116, 18 Shimasaku et al. 2006

log (number/Mpc 3/log lum)

Hu E. M., Cowie L. L. and McMahon R. G. 1998,

ApJL 502, L99 –3
Meinköhn E. and Richling S. 2002, A&A 392, 827
Noll S. et al. 2004, A&A 418, 885
Partridge R. B. and Peebles P. J. E. 1967, ApJ 147,
868
Shapley A. E. et al. 2003, ApJ 588, 65 –4
Shimasaku K. et al. 2006, PASJ 58, 313
Tapken C. et al. 2004, A&A 416, L1
Tapken C. et al. 2006, A&A 455, 145
Tapken C. et al. 2007, A&A 467, 63
–5
35 35.5 36 36.5 37
log Lyα luminosity [W]

The Messenger 128 – June 2007 55

Astronomical News

Photo: H. H. Heyer, ESO

Atacama landscape
view near the
ALMA Site Museum at
3 200 m.
Astronomical News

Nature Around the ALMA Site – Part 2

Michel Grenon Valeriana nivalis

(Geneva Observatory, Switzerland) Festuca chrysophylla
Adesmia spinosissima
Baccharis tola
The natural environment around the Parastrephia quadrangularis
ALMA site, its flora, fauna and land- Fabiana denudata
scape morphology, are presented and Junellia seriphioides
interpreted in terms of combined geo- Trichocereus atacamensis
logical and climatic evolution with, in Atriplex imbricata
Opuntia camachoi
parallel, the necessary biological adap-
tions. This part covers vegetation and
animal life.
uVg
Sairecabur
Mulinum crassifolium
Deyeuxia curvala
Pycnophyllum bryo
oVg
Figure 1: The vegetation
belts as described
m by Richter (2003) on the
slopes of Sairecabur
5 500 volcano, NW of ALMA,
with indication of char-
acteristic plants.
5 000
4 500
4 000
3 500
3 000

The vegetation belts

Figure 2: The Maihuen-
iopsis pillows show the
The vegetation belts contain plant asso-
effect of the katabatic
cations adapted to a given range of at- winds. The side exposed
mospheric parameters, soil composition to the cold flow is red-
and texture. A recent phytosociological der, an indication of a
higher content in antho-
survey by Richter (2003) provides vege-
cyanine, a molecule
tation transects in the ALMA area, namely acting as anti-freeze.
at Sairecabur, Toco Toco and Miñiques
(Figure 1). On the western slopes of Toco
Toco, all belts are present, because the
rocks are old enough to have been al-
tered into sand and clay, and hence are
able to retain water.

At the lowest level, 2 900–3 350 m, the

limiting factors are the scarcity of precip-
itation, the high evapotranspiration and
the salt and nitrate content in the soil.
Atriplex communities – tall greyish bushes Figure 3: Thanks to the
reduction of the leaf
– develop on the shore of the Atacama
size, Fabiana bryoides
Salar in salt rich soils. On undisturbed resists dessicating
rocky places, the cactus Maihueniopsis winds and may grow up
camachoi forms colonies of spiny pillows to 4 900 m.
(Figure 2), among several other plants.
In wind shadow places, the tall cactus
Trichocereus atacamensis, an invader
from Argentina, grows.

At higher altitudes, evapotranspiration

becomes the dominant factor. The effi-
ciency of the evaporation increases with
the square of the wind velocity. Low
humidity and high wind velocity favour
high rates of water loss by the plant
epiderm. Several techniques are devel-
oped by plants to reduce these losses.
In the second vegetation zone, 3 350–
3 850 m, is found Fabiana bryoides,
with leaves reduced to minute rosettes,
forming a compact cover on the stem,
mimicking coral branches (Figure 3). In
the same zone, Fabiana denudata repre-
sents an extreme case: its leaves are
suppressed and photosynthesis takes
place at the upper stem surface. The
whole plant is varnished with a viscous
resin. In other plants, such as some
Senecios, the water losses are minimised
through the development of a white to-
mentum of dense crispy hairs, reflecting
the solar radiation and setting the wind
velocity to zero at the epiderm surface.
In the zone 3 850–4 300 m, the vegeta-
tion becomes scarce and the grami-
naceae herbs are dominant. The golden
Festuca chrysophylla gives the landscape
its colour and specific character (Figure 4).
Behind the tufts of Festuca, in the wind
shadow, several plant species may de-
velop. Parastrephia quadrangularis is an-
other typical plant with leaves reduced

The Messenger 128 – June 2007 57

Astronomical News Grenon M., Nature Around the ALMA Site – Part 2

to scales, covering the stem as tiles. The Figure 4: Festuca

chrysophylla forms large
association of Festuca, Parastrephia and
populations giving a
Baccharis provides the preferred pasture golden aspect to the
of guanacos and vicuñas (Figure 5). landscape at altitudes
where most bushes, ex-
cept Parastrephia, have
disappeared.
The high-altitude vegetation

In the zone 4 350 m to 4 850–5150 m, the

wind intensity, the eolian erosion and the
temperature are the limiting factors for
vegetation. On flat surfaces, plants have
to minimise their cross section with re-
spect to the thermal and zonal winds. To
expand horizontally is a frequently en-
countered adaptation, e.g. by Pycno-
phyllum bryoides (Figure 6) and by the
Calyceras genus (Figure 7). This strategy
is also adopted by some dwarf trees Figure 5: Vicuñas in
the altoandine steppe
as Adesmia sp., which develops an under-
near Maricunga Salar,
ground stem and branch systems ex- 3 800 m.
tending well below the surface. Leaves
are covered with hygroscopic glanduli-
ferous hairs able to absorb directly the
humidity from the air.

In rocky places, plants may develop pro-

vided the soil is evolved. Fresh lava or Figure 6 (below):
lapilli cannot retain water close to the sur- Pycnophyllum bryoides
face and several centuries of weathering expands as rings on
flat gravel surfaces (left).
are needed before the first plant colony Leaves are reduced
may settle. On older substrates, such as to ovoid cones (right),
at Toco Toco, plants use rock cracks, petals are translucid,
at wind shadow, to expand their roots in and so the plant offers
a minimum surface to
clay, searching for residual humidity. dessicating winds.
Nototriche holosericea and Chaetanthera
revoluta (Figure 9), and Oxalis sp. (Fig-
ure 10), are typical examples of this be-
haviour.

In small valleys oriented NS, perpendicu-

lar to the afternoon and zonal winds,
the soil and plant evaporation is notice-
ably reduced. Snow may accumulate
during winter and stay until the next blos-
soming season. Plant communities re-
quiring less protection against evapotran-
spiration may develop up to very high
altitudes, as the Werneria-Senecio asso-
ciation, consisting of a dozen different
species. Around ALMA, this association
is characterised by the presence of
Werneria poposa and Valeriana nivalis
(Figure 11). In these sites, the limiting
factor is the temperature, which reduces
the duration of the vegetation period to
nearly zero above 4 850–4 900 m.

58 The Messenger 128 – June 2007

 Figure 8: This Adesmia
sp. is an underground
dwarf tree. Strong
spines prevent grazing
by viscachas or vicuñas
(Toco Toco, 4 500 m).

Figure 7 (above): In Figure 9 (below): In Nototriche holosericea (left)

Calyceras genus, the leaves are undulated and cerebriform; the surface
hemispherical head of available for gas exchanges exposed to the wind
green flowers is the only is manyfold. A white indumentum protects the leaf
part emerging above from transpiration and UV radiation. Chaetanthera
ground level (Laguna revoluta flowers (right) open on top of 1 cm long
Miscanti, 4 000 m). cones, covered by hairy leaves (Llano de Pajonales,
4 410 m).

Figure 10: This Oxalis

sp. expands along
narrow cracks, in sunny
places, at Toco Toco,
4 400 m.

The Messenger 128 – June 2007 59

Astronomical News
Figure 11: Werneria poposa (left) grows in wet and
wind-shadow rocky places, close to Valeriana nivalis
(right), a rather common plant reaching the upper
limit of the vegetation; its thick root is strongly aro-
matic, reminiscent of the Celtic Nard (N of Toco
Toco, 4 520 m).
Figure 12: Four representative plants at the vege-
tation limit in humid places at Cerro Toco Toco.
Top left: Calandrinia sp.; top right: Menonvillea sp.
(4 820 m); lower left: Perezia atacamensis (4 700 m);
lower right: Werneria pinnatifida (4 700 m).
A notable character of the high-altitude
flowers is the restriction of their colour
range. They look unattractive to the
human eye: blue, orange, red colours are
no longer present. Most flowers are
white, yellowish or at best bright yellow
(Figure 12). With a very low ground
coverage, plants appear to rely more on
petal UV-reflectivity, (increasing the con-
trast between the flowers and the
60 The Messenger 128 – June 2007
Grenon M., Nature Around the ALMA Site – Part 2
ground), than on bright colours, in order
to attract hymenopter pollinators whose
eyes, or ocelli, are sensitive down to
UV-B radiation.
The present geographical distribution of
high Andes plants reflects the restric-
tion of areas consequent on the climate
warming after the Ice Age. Intermediate
altitude plant communities had to move
up by about 1 km, migrating towards
the altiplano on gentle slopes east of the
Atacama Desert core, or towards the
top of isolated mountain ranges. Those
already growing on the altiplano during
the Ice Age are found presently on iso-
lated peaks, close to the upper vegeta-
Figure 13: Calceolaria stellariifolia is a rare plant
found between 4 000 and 4 300 m around ALMA. Its
geographical distribution shows a typical area
disjunction consequent on the postglacial climate
warming.
Figure 14: Senecio aff. algens grows between dark
rocks exposed to the west. The altitude of Toco Toco
station, at 5150 m, is the highest for flowering plants
observed so far near ALMA.
tion limit. As a result, the areas occupied
by high Andian plants are now disjunct.
An example of severe area restriction is
that of Calceolaria stellariifolia (Figure 13),
found in only half a dozen sites, spread
over 1500 km in the high Andes. When
isolated, plants may follow divergent ge-
netic evolution, they have no chance to
merge their genes again before the next
glaciation.
Plants reaching the maximum altitude
around ALMA belong to the Senecio
genus. Senecio Puchii is frequently seen
up to 4 750 m. Senecio aff. algens (Fig-
ure 14), replaces it at higher altitude, in
sunny places between 4 850 and 5150 m.
Hot springs and high altitude vegas Figure 15: High andine
vegas are formed of
compact material, strong
Wet biotopes are due either to hot springs enough to resist the
in hydrothermal fields as at El Tatio, or weight of vicuñas or
to the development of high altitude vegas, even llamas (Chungara,
4 500 m).
the southern counterpart of peat bog in
the Northern Hemisphere. In the Andes,
the bog is made up of a very compact
plant interlacing, where the Juncacea
Oxychloe andina plays a central role (Fig-
ure 15). Water pools are resting places
for many birds, including the large andine
goose, as well as for migrating birds,
transporting seeds from one vega to the
next. The vega vegetation is hence very
uniform along the Andes. Hot springs Figure 16: Frankenia
triandra grows on warm
host highly specialised life forms, adapted
bumps in the immediate
both to high soil temperature and high vicinity of hot springs
mineral content, such as Frankenia trian- (El Tatio geyser, 4 350 m).
dra (Figure 16), widespread over 1200 km
in the Central Andes.

Many more adaptations

Not only plants, but also wildlife, show

surprising adaptations to the extreme arid
and cold conditions prevailing in North
East Chile. As an example, a group of
reptiles from the Iguanidae family, the
genus Liolaemus, has evolved in parallel
to plants, into an incredible number of
species and varieties, able to colonise
nearly all biotopes from the margin of the
absolute desert up to the upper vege-
tation limit. As insects are often uncom-
mon, they had to become omnivorous,
some totally vegetarian. At high altitude,
they are adapted to absorb the solar
radiation, and hence able to move even if
the external temperature is well below
0˚ C. The specimen shown in Figure 17 is
an extreme case of this kind of adapta-
tion.
... and final remarks
After this quick overview of the ALMA site
natural history and climate, and of some
of the adaptations imposed on life – at
the cost of massive extinction – we may
appreciate even more the apparent para-
dox that scientists had to put the ALMA
complex at the upper limit for evolved
terrestrial life, to be in a good position to
search for molecules in space, progeni-
tors of terrestrial and extraterrestrial life.
We may simply regret that, thanks to ser-
vice observations, very few observers
will have the opportunity to visit and enjoy
the unique natural surroundings of Chaj-
nantor, even if Homo sapiens is still not
yet perfectly tuned to activities above
5 000 m. A more serious concern is the
expected final extinction of a large frac-
tion of the high-altitude species as a re-
sult of the global warming during the
XXIst century. A local warming by about
+ 3.5˚ C would increase the vegetation
belts altitude by 500 to 600 m.
All meteorological, climatic and biological
indicators converge to qualify the Chaj-
nantor site as the best possible for milli-
Figure 17: This Lio-
laemus sp. lizard is per-
fectly mimetic with
the rocky environment.
Able to absorb the solar
radiation, it may start
hunting in the early morn-
ing when the air tem-
perature is still – 25˚ C
(Laguna Verde, East of
Copiapo, 4 500 m).
metre and submillimetre astronomy by
the end of the XXth century. Whether its
exceptional qualities will be preserved
through the XXIst century, depends on
the amplitude of the climate change and
the evolution of precipitation patterns in
the area. Most Global Circulation Models
(GCM) predict an aridification of the Ata-
cama area, but an increase of wetness
and precipitation on the eastern side of
the Andes.
Reference
Richter M. 2003, Lyonia 4(1), 1

The Messenger 128 – June 2007 61

Astronomical News
Using the h-index to Explore the Scientific Impact
of the VLT
Uta Grothkopf, Claudio Melo,
Christopher Erdmann, Andreas Kaufer,
Bruno Leibundgut (all ESO)
The productivity and scientific impact of
observatories and individual instru-
ments are one measure of their suc-
cess. This article presents the results of
a study where we have applied the
h-index, previously proposed for indi-
vidual researchers, to major ground-
based observatories (VLT, Keck, Gemini,
Subaru) as well as individual VLT instru-
ments. The concept is expanded by
exploring the time-dependence of the
h-index h(t). Overall, the VLT appears
to be among the most successful 8-m-
class telescopes. We also show that
ESO instruments are making important
contributions to progress in astronomy.
Introduction
In order to examine their return on invest-
ment, major observatories around the
world have developed metrics to trace
their scientific output. Such metrics often
focus on the observatories’ productivity
and impact in the scientific community.
These two factors are typically measured
through the number of scientific publica-
tions based on astronomical data and the
citations these publications generate, re-
spectively. The methodologies used to
compile the ESO Telescope Bibliography,
a database that lists all papers based on
ESO data, as well as publication and
citation statistics derived from this data-
base have been documented in Leib-
undgut, Grothkopf and Treumann (2003)
and Grothkopf et al. (2005). The Tele-
scope Bibliography is publicly available
via the web (http://www.eso.org/libraries/
telbib.html) as well as through the “Select
References In: ESO/Telescopes” filter
at the ADS (see Delmotte et al. 2005 for
more details).
In a recent paper, Hirsch (2005) proposed
a new index to measure research output,
the so-called h-index. While originally
meant to analyse the productivity of indi-
vidual researchers, we recently intro-
duced it into telescope statistics (Groth-
kopf and Stevens-Rayburn 2007). Based
on h, we developed h(t) which reflects
changes of the h-index in the course of
62 The Messenger 128 – June 2007
time. In this paper, we apply h and h(t) to
selected observatories as well as to the
VLT instruments.
There are several caveats to bibliometric
studies, in particular when used for
comparison across various institutes. De-
spite attempts to synchronise the meth-
ods applied to compile science biblio-
graphies, criteria for paper selection are
still defined by the individual observato-
ries and are therefore not identical. Simi-
larly, methodologies for building tele-
scope bibliographies vary (for instance,
retrieval of relevant publications through
database (ADS) searches alone, through
screening of paper journals, etc.). Even
more importantly, comparing telescope
statistics is problematic because of the
different features and ways of operation
of ground-based and space-based tele-
scopes, different apertures and numbers
of instruments, different wavelengths
of observations, etc. Any comparison has
therefore to be interpreted with utmost
care in order to avoid unbalanced or
wrong conclusions.
Use of bibliometrics by observatories
In preparation for a presentation given at
the IAU General Assembly in Prague,
Czech Republic in August 2006, we con-
ducted a survey among large observa-
tories to better understand what kind of
telescope statistics are compiled at ob-
servatories, at which intervals, and by
whom.1 All respondents regularly gather
total and/or average numbers of their
refereed publications, the majority also
monitor unrefereed publications (e.g.
conference proceedings), either regularly
or on request. Citation statistics of refer-
eed publications are collected by all re-
spondents, even though only half of them
does so on a regular basis, the remaining
50 % only on request.
As can be expected, observatories com-
pile statistics tailored to their individual
needs, for instance number of publica-
tions and citations per instrument, pro-
1
 he following observatories responded to our
T statistics. In order to determine this value,
questionnaire (http://www.eso.org/libraries/
telstats-questionnaire.html ): CFHT, Chandra,
Gemini, HST, Isaac Newton Group, Keck, NRAO,
Subaru, XMM Newton. For more information,
contact Uta Grothkopf at esolib@eso.org.
gramme type, observing cycle, observ-
ing mode, etc. In addition, almost all
respondents investigate their high-impact
papers (“most productive instrument,
programme, individual authors”, etc.) A
study of highly cited papers and their
distribution among facilities is carried out
every year in April at the Space Tele-
scope Science Institute (Madrid and
Macchetto 2007).
All methodologies used by observatories
have some advantages and some dis-
advantages: counting publications meas-
ures the productivity of facilities, but does
not indicate whether or not these publi-
cations actually have any influence on the
advancement of astronomy. Looking at
the numbers of citations does indicate
the impact among the astronomical com-
munity, but values can easily be inflat-
ed by a few extremely highly cited pa-
pers. Investigating so-called high-impact
papers and their distribution across ob-
servatories is less biased, but retrieving
such statistics is by far not as straight-
forward as mere publication and citation
counting.
Several in-depth studies have been car-
ried out by Trimble et al. who investi-
gated productivity and impact of optical,
space-based, and radio telescopes,
respectively (Trimble, Zaich and Bosler,
2005, 2006; Trimble and Zaich 2006).
The authors analyse papers, citations
and impact factors of articles from
18 journals regarding their distribution
among facilities, thus avoiding the bias
that typically can be noted in studies
from individual observatories.
The h-index
In order to overcome some of these dis-
advantages, J. E. Hirsch of the University
of California at San Diego suggested
a new and surprisingly simple measure
which he called the h-index (Hirsch
2005). While the h-index originally was
meant to be a measure for the research
output of individual scientists, we ex-
tend it here to observatory publication
one needs a list of all relevant papers,
ranked by decreasing citation counts; h
can then be found where the citation
count is at least as high as the rank. Thus,
the h-index combines measure of pro- 200 Figure 1: The h-index can be found
where the counted publication number
ductivity (number of publications) and
equals the number of citations.
impact (average number of citations per Neither a large number of publications
paper) and is therefore more balanced with low citation rates (right edge of
than most other measures used for bibli- 150 the x-axis) nor individual papers with
extraordinarily high numbers of
ometric studies (Figure 1).
citations (upper edge of the y-axis) will
alter h, therefore this value is more
Instead of measuring the research output Citations
balanced than many other bibliometric
of individual researchers, the concept 100
measures.
can also be applied to entire observato-
ries or specific observing facilities –
always bearing in mind the caveats de-
scribed above. In a recent paper, we
50
introduced the h-index into telescope sta- h-index
tistics by computing it for selected pub-
lication years of some major observato-
ries (Grothkopf and Stevens-Rayburn
2007). 0
0 50 100 150 200
Publication number

The m-parameter
Observatory Range of years Years since h m Table 1: Range of years of publica-
of publications first publication tions, number of years since first
It must be noted that, tempting as it may
VLT 1999–2006 8 79 9.9 publication, as well as h and m of the
be, comparing h alone among observ- observatories included in our study.
Keck 1996–2005 11 113 10.3
atories (or researchers) does not lead to a
Gemini 2000–2006 7 33 4.7
meaningful result. Facilities that have
Subaru 2000–2006 7 41 5.9
been operative for many years obviously
had much more time to produce publi-
cations and accumulate citations, hence
their h-index can be expected to be con- h-index Figure 2: h-index for the VLT, Keck,
Gemini, and Subaru observatories as
siderably higher than that of younger 120
of April 2007. Note that h correlates
facilities. Hirsch therefore introduced the with the number of years of operation.
so-called m-parameter. For individual re- Hence, older observatories tend to
100
searchers, this value is computed by have higher h indices.
dividing h by the numbers of years since
publication of the first paper. Correspond- 80
ingly, when applied to observatories, h
is divided by the number of years of oper-
ation; hence it reflects the various ‘life- 60
times’ of facilities.
40

Applying h and m to observatories

We present here h and m values for the
VLT, Keck, Gemini and Subaru observa-
tories. In order to compute h, we ob-
tained bibcodes of all papers pertaining
to the observatories’ publication lists in
the following way: for the VLT, bibcodes
are stored in the ESO Telescope Bibli-
ography (http://www.eso.org/libraries/
telbib.html ). References of Keck and
Subaru papers were retrieved from the
web (Keck Science Bibliography at http://
www2.keck.hawaii.edu/library/keck_
papers.html and Publishing Results from
Subaru, http://www.naoj.org/Observing/
Proposals/Publish/index.html ) and were
20
0
VLT Keck
1999–2006 1996–2006
translated into bibcodes. Papers using
Gemini data were found using the “Select
References In” filter on the ADS main
search screen. These bibliographies in-
clude only refereed papers. The respec-
tive ranges of publication years are
shown in Table 1. For uniformity, we end
the range of publication years for all ob-
Gemini Subaru
2000–2006 2000–2006
servatories in 2006. Citation counts for all
publications were obtained from the ADS
as of April 2007.
For each observatory, citation counts
were then ranked in descending order,
and h was computed. Figure 2 shows the
h-indices of the respective observatories.

The Messenger 128 – June 2007 63

Astronomical News Grothkopf U. et al., Using the h-index to Explore the Scientific Impact of the VLT

We also computed the m-parameter by Parameter m Figure 3: Parameter m

of the observatories
dividing h by the specific number of 12
included in this study (as
years since the first publication (see Ta- of April 2007).
ble 1). Resulting values are given in Fig-
10
ure 3. This brings the VLT and Keck
to the top of the list, and also the young
facilities Gemini and Subaru perform well. 8

One should note that both h and m de-

pend on the number of telescopes per fa- 6
cility. A facility like the VLT with four tele-
scopes will produce more papers more
4
quickly and hence the h-index will in-
crease faster, although it is not obvious
exactly how much of an effect this is. 2
For a facility like the VLT, with telescopes
becoming operational over some time,
it will be difficult to quantify this effect in 0
detail. VLT Keck Gemini Subaru
1999–2006 1996–2006 2000–2006 2000–2006

The h-index versus time

Although the m-parameter reflects the

total number of years of operation of h-index
a facility, both h and m are integrated 120
values that don’t show how these values
were achieved in the course of time.
100
We have therefore further developed the
h-index to analyse its evolution, h(t).
80
To compute h(t), we obtained the citation
history for each paper included in our
study from the ADS. The citation history 60
shows how many citations a paper gener-
ated in a given year. For instance, a paper
40
published in 2002 will generate x citations
in 2002, y citations in 2003, z citations
in 2004, etc. In order to calculate h for a 20
year Y, we add up all citations up to the
year Y, using the citation history of each
paper. The papers are then listed in de- 0
creasing order, and the h-index for that 0 1 2 3 4 5 6 7 8 9 10
year Y is computed. Years after publication

Figure 4 shows h over time applied to the
observatories in our study. Trends in-
dicated by h and m are confirmed. Keck
has always been performing extremely
well, with a steeply increasing h-index
right from the start. The VLT started well
and has even improved during recent
years. Both Subaru and Gemini are now
on their way up after a slightly slower
start.
In this study, all observatories are treated
as entities. This does not accurately re-
flect their early years. Keck I and II,
Figure 4: h(t) of Keck, VLT, Subaru and Gemini by
years after first publication (as of April 2007).
Gemini North and South as well as VLT
UT1, 2, 3, and 4 all came online sequen-
tially, increasing the available obser-
ving time over the years. A future study
may investigate h(t) based on the actual
observing time of each observatory.
Keck Subaru
VLT Gemini
Performance of the VLT instruments
In order to investigate the specific perfor-
mance of the VLT instruments, we ap-
plied h, m, and h(t) to them. We restrict
this analysis to the VLT because instru-
ment-level information for La Silla papers
has been systematically recorded only

64 The Messenger 128 – June 2007

starting in 2002. The results are presented Instrument h-index Years since m Number Total number Ratio
first publication of papers of citations
in Table 2. Although the absolute num-
VLT
ber of papers and citations differ, the very
FORS1 56 8 7.0 525 14 267 27.2
first instruments of the VLT (FORS1,
ISAAC 49 8 6.1 413 9 308 22.5
ISAAC, UVES, FORS2) have a compara-
UVES 48 7 6.9 474 9 326 19.7
ble performance. As mentioned above,
FORS2 46 7 6.6 342 9 006 26.3
the number of years in operation is
NACO 22 5 4.4 111 1942 17.5
an important factor. Young instruments
FLAMES 14 4 3.5 67 877 13.1
necessarily have a lower h-index. This
VIMOS 14 3 4.7 44 623 14.2
can be partially compensated for by
VLTI
normalising the h-index by comparing
VINCI 2 14 5 2.8 38 572 15.1
parameter m.
MIDI 7 3 2.3 19 226 11.9

As done for the observatories, a better

glimpse of how the instruments are per- Table 2: Performance of VLT and VLTI instruments 2
 ote that VINCI was offered to the
N
that have been operative for two or more years and community only from October 2002
forming can be achieved by measuring
produced at least ten papers. to October 2003.
the h-index versus time. The results of h(t)
as of April 2007 for VLT instruments are
shown in Figure 5 relative to the year of
the first paper of each instrument. In this
graph, we included only VLT instruments
that have been operative for two or more
years and produced at least ten papers.

An obvious result is that the first-genera- h(t)

tion VLT instruments (FORS1, UVES,
ISAAC, FORS2) are alive and kicking with
no sign of changing their slopes. They
clearly set a very high standard for future
instruments.
With respect to the young instruments,
we can say that they are … young! Their
h-indices are growing at a slow pace. Be-
sides this slow increase of the h-index
for the young instruments, can we predict
whether they will look like their older sib-
lings in future?
We see from Figure 5 that most of the
young instruments have a much harder
time to take off than FORS1 and UVES,
but they are not particularly different from
the h(t) curves of FORS2 and ISAAC.
But what is behind the plot shown in Fig-
ure 5? Does the h(t) have a practical use
in judging instrument performance?
Can it help observatories to decide wheth-
er a given instrument needs to be up-
graded, improved or even decommis-
sioned? The answer is not simple and the
reader should take Figure 5 with caution.
As defined above, the h-index depends
on the number of papers and their ci-
tation counts. These two quantities are
certainly instrument-dependent. For
example, detector sensitivity, spectral or
spatial resolution, wavelength coverage,
60
50
40
30
20
UVES
10 FORS1
FORS2
ISAAC
NACO
0 VIMOS
0 1 2 3 4 5 6 7 FLAMES
Years after first publication
Figure 5: h(t) of selected VLT instruments by years
after first publication (as of April 2007).
quality of the documentation and data- h. As time goes by the community gets
reduction software (pipelines). These are used to the modes, pipeline, etc. At this
all factors that play a role in the way h(t) time the number of papers and citations
evolves. increases fast, and so does the h. Years
later, all possible (hot) applications start
The intrinsic nature of the instrument it- to run out and the papers gather fewer
self has an influence on its h(t). All in- citations. New (and hopefully better) in-
struments have a difficult start with a low struments come. At this point the h-index

The Messenger 128 – June 2007 65

Astronomical News
keeps growing, but at a much slower
pace. Thus the h(t) would have a curve-
of-growth shape (similar to those found in
chemical abundances studies).
In the case of instruments designed for
individual observations, papers can
be based on a few observations and the
full cycle (from proposal submission to
paper acceptance) can be as short as
one year. In contrast, papers presenting
results based on a large amount of
data collected by survey-like instruments
(e.g., VIMOS and FLAMES) can take
years to be released. But when these pa-
pers are finally published, the impact can
be immense.
Another bias hidden behind the compu-
tation of h(t) has to do with the com-
plexity of the new instruments, which in
some cases are aimed at tackling prob-
lems that require the use of cutting-
edge technology. Using a complex (new)
technology usually implies large over-
heads. This is the case of AO instruments
and even more pronounced in the VLTI
observations whose complexity is of a
higher degree than that of any other VLT
instrument.
However, the h(t) can also be influenced
by circumstances not related to the in-
strument at all. For instance, there is no
question that one of the strengths of
the first suite of VLT instruments (FORS1,
UVES, ISAAC and FORS2) is its versatil-
ity, being used from Solar System to cos-
mological applications. In addition to
this versatility, these instruments were fa-
voured by the fact that they were the
only ones available to the community dur-
ing the first years of operation of the VLT.
Moreover, they were mounted almost ex-
clusively on each of the UTs. Nowadays,
the situation is very different. New instru-
ments have to compete for UT time.
For instance, on Kueyen UT2, the time is
shared between UVES, FORS1 and
FLAMES. Therefore, in order to be fair, the
h(t) curve shown in Figure 5 has to be
corrected by the effective fraction of time
used by each instrument.
66 The Messenger 128 – June 2007
Grothkopf U. et al., Using the h-index to Explore the Scientific Impact of the VLT
The bottom line is that we should take the
results presented in Figure 5 with cau-
tion since none of the potential biases
discussed above were taken into account
and therefore the results presented here
are very preliminary. Having said that,
it might be interesting to peruse the ef-
fort of using the h-index to measure in-
strument impact and performance in a
deeper and more thoughtful study where
all biases are considered.
Conclusions
The h-index is a simple, yet powerful in-
dicator that has some important ad-
vantages compared to other bibliometric
methods:
– it combines productivity and impact;
– it can be relatively easily determined
using the ADS;
– it is neither affected by a large number
of publications with few or no citations
(which usually suggests high productiv-
ity, but not necessarily high impact) nor
by extraordinarily high citations of only
a few papers (which inflates citation
counts).
The h-index is therefore more balanced
than other measures if one bears in
mind the usual caveats intrinsic to biblio-
metric comparisons. It is important to
note that h depends on the number of
years of operation and therefore needs to
be combined with the m-parameter
in order to avoid biased interpretations.
Hirsch (2005) points out that the h-index
is prone to depend on the field of study
(Physics, Biology, etc.). This is true even
within astronomy. For instance, astrono-
mers in the gamma-ray burst community
have higher h-indices than their col-
leagues in other areas, partly because of
the importance of the field, but also be-
cause of the size of the collaborations
and the rate of publications.
If this effect is present with regard to in-
struments, it may indicate that a given
instrument is flexible enough to produce
papers in a wide range of astronomical
fields. The numbers in Table 2 show that
the first-generation VLT instruments have
similar h-indices. They constitute an
example of such versatility, since they are
used for observations ranging from the
Solar System to cosmological applica-
tions. However, other biases influencing
h(t) need to be carefully corrected before
a detailed comparison is made. In this
sense, our results concerning the h-index
for instruments are regarded as prelimi-
nary.
Acknowledgements
This research has made extensive use of NASA’s
Astrophysics Data System Bibliographic Services.
References
Delmotte N. et al. 2005, The Messenger 119, 50
Grothkopf U. et al. 2005, The Messenger 119, 45
Grothkopf U. and Stevens-Rayburn S. 2007, in
“Library and Information Services in Astrono-
my V”, eds. S. Ricketts, C. Birdie and E. Isaksson,
ASP Conference Series, astro-ph/0610274
Hirsch J. E. 2005, Proc. Nat. Acad. Sci. 102, 16569
Leibundgut B., Grothkopf U. and Treumann A.
2003, The Messenger 114, 46
Madrid J. P. and Macchetto F. D. 2007, in “Library
and Information Services in Astronomy V”, eds.
S. Ricketts, C. Birdie and E. Isaksson, ASP Con-
ference Series
Trimble V., Zaich P. and Bosler T. 2005, PASP 117,
111
Trimble V., Zaich P. and Bosler T. 2006, PASP 118,
651
Trimble V. and Zaich P. 2006, PASP 118, 933
Astronomical News
Status of Women at ESO: a Pilot Study
on ESO Staff Gender Distribution
Francesca Primas (ESO)
Equal career opportunities require
working conditions that make it possi-
ble to reconcile family needs and career
development. This article describes
the goals and main findings of a pilot
investigation that has recently been
­carried out at ESO focusing on gender
balance issues.
Over the past decades, several studies in
Europe (e.g. within the FP5-FP6 pro-
grammes) and the United States have
considered gender distribution, and in
particular the status of women, in dif-
ferent sciences. Gender equality and dual
careers are just two of the many aspects
that have been closely scrutinised. In
astronomy, the Baltimore Charter and the
Pasadena Recommendations are among
the main outcomes of these investiga-
tions and provide guidelines on how the
situation of women in astronomy could
be improved. Further, several working
groups have been established by interna-
tional scientific bodies. With ESO as a
multi-cultural research organisation, it be-
came clear that a similar, systematic study
of its current gender distribution was an
important goal to achieve. This article
presents the first results of such a study.
The project and its goals
The “Status of Women at ESO” is a pro-
ject that, as a start, aimed at evaluating
the current gender distribution at ESO,
thus providing a snapshot of the status of
women at ESO. This task necessitated a
systematic and impartial breakdown of
the ESO organisational structure in terms
of female versus male personnel mem-
bers by division and rank over a given
time period. The team was chaired by
Francesca Primas and included Roland
Block, Mark Casali, Nathalie Kastelyn,
Rubina Kotak and Bruno Leibundgut. We
focused on the period 1999–2006, which
has witnessed a significant expansion
in the number of staff members owing to
the start of VLT operations, and on the
following groups:
a) ESO International Staff Members
and Paid Associates by division and by
rank;
b) ESO Faculty members, Fellows and
Students;
c) ESO Governing Bodies and Commit-
tees: Council, Observing Programmes
Committee, Scientific and Technical
Committee, Users Committee, Finance
Committee;
d) ESO Visitors;
e) invited speakers at ESO Lunch Talks
and the Munich Joint Astronomy Col-
loquia.
The project and its outcomes
Once all the data were gathered, we
compiled and analysed statistics on the
distribution of female and male employ-
ees by various categories. The numbers
reported here cover all international staff
of ESO, i.e. at both the ESO Headquar-
ters in Garching and ESO Chile, with the
latter including the Vitacura offices and
the La Silla Paranal observatory. Where
appropriate, we have split the analysis for
the two main ESO regions. In the follow-
ing sections, we will briefly comment on
the results obtained for the different cate-
gories (a–e) that were scrutinised.
Figure 1: Yearly gender distribution of all ESO
International Staff Members and Paid Associates
(dark blue: male, light blue: female).
1999
2000
2001
2002
2003
2004
2005
2006
10 20 30 40
%
a) ESO staff members
Established members of ESO personnel
are identified as International Staff Mem-
bers (ISM), whereas Paid Associates (PA)
belong to the so-called non-established
personnel. Typically about 7 to 10 PA are
employed at ESO per year and these are
included in the overall statistics of ESO
staff.
Figure 1 displays the fraction of female
ESO staff over the past eight years. Dur-
ing the same time the ESO staff in-
creased from 239 ISM in 1999 to 329 ISM
in 2006. This is also reflected in the
recruitment statistics: ESO recruited
33 women (17.7 %) compared to 153 men
(82.3 %) between 2000 and 2005. Al-
though the number of women employed
as ISM has increased from 44 to 59
over this period, it has clearly lagged
behind the pace of male hiring. Among
the women hired, nearly half (16 or
48.5 %) were scientists, 12 (36.4 %) were
administrative staff and only 5 women
(15.2 %) were engineers. When the break-
down is done by ESO division, one finds
that while the Office of the Director Gen-
eral and Administration have a nearly
even gender balance, the technical divi-
sions have, in general, a very low frac-
50 60 70 80 90
The Messenger 128 – June 2007 67
Astronomical News
tion of women on their staff. The fraction
of female Paid Associates is slightly
higher than that of ISM; this also holds in
the more technically-oriented divisions.
When the grade is taken into account, the
difference becomes even more striking:
during the past three years, the percent-
age of female staff (ISM + PA) has re-
mained rather stable, around 50–60 % for
salary grades 5–7, but only around 10 %
for grades 8–11. This continues further
into the management level (grades 12 and
above) where at the moment there are
only two women. Women are clearly
under-represented at the higher eche-
lons, presumably corresponding to posi-
tions with a higher scientific and techni-
cal profile. It should be noted that the
current Director General and one division
head are women. By total numbers, ESO
has more than doubled the number of
women in grades 8 to 11 (from 16 to 34),
but the relative increase remains rather
small (only 4 %).
b) The ESO Faculty, Fellows and Students
The ESO Astronomer Charter defines the
roles of astronomers and their member-
ship in the ESO Faculty. As of 2006, there
are 17 women in the Faculty (to be com-
pared to 74 men), and their fraction has
increased fairly steadily over the past de-
cade.
However, we note that the highest frac-
tion of female Faculty members is cur-
rently at the Assistant Astronomer level,
the entry level for junior staff. The Sci-
entist level (astronomers with reduced
research time), also have a higher fraction
of women, as can be seen from Figure 2.
The global average, including all levels
(Full, Associate, Assistant and Scientist) is
18.7 % of women.
Complete statistics for ESO fellows and
students are available from the year 2000
onwards for both ESO sites, including
the gender distribution among the appli-
cants and the gender distribution in the
Fellowship and Studentship Selection
Committees (FSSC). Table 1 shows the
total number and percentages of all ESO
Fellows and Students during the last
seven years.
68 The Messenger 128 – June 2007
Primas F., Status of Women at ESO: a Pilot Study on ESO Staff Gender Distribution
Full
Associate
Assistant
Scientist
10 20 30
total number
Figure 2: Gender distribution (by total number) of
the three categories of ESO Faculty members, and
Scientist, in the year 2006 (dark blue: male, light blue:
female).
% Figure 3: Yearly gender distribution
100
80
60
40
20
0 Female Applicants
2001 2002 2003 2004 2005
The percentage of females among the
selected fellows has been between
10 and 40 % over the past few years. The
comparison with the distribution among
the applicants (Figure 3) shows that the
gender balance is reasonably maintained,
i.e. there is no obvious discrimination
against female candidates.
However, the years 2003 and 2004 show
that this is not universally true. In 2004,
for instance, only one female fellow was
selected out of 34 female applicants!
Also, it should be noted that the fraction
of female fellows is significantly higher
than for female members in the Faculty.
Compared to the Assistant Astronomer
level in the Faculty, the gender distribu-
tion is nearly the same. However, despite
the increased number of Fellowships
offered in Chile for Paranal, we have ob-
served a steady decrease in the number
of female applicants over the last few
Female | Male
Fellows Students
Garching 10 | 34 22 | 29
Chile 15 | 36 11 | 17
All 25 | 70 33 | 46
Garching (%) 22.7 | 77.3 43.1 | 56.9
Chile (%) 29.4 | 70.6 39.3 | 60.7
All (%) 26.3 | 73.7 41.8 | 58.2
Table 1: Total number and percentages of
ESO fellows and students (2000–2006).
among the selected ESO fellows and
applicants to the ESO Fellowship,
over the period 2000–2006 (for both
Garching and Santiago).
Male Applicants
Male Fellows
2006 Female Fellows
years, which has stopped only in the past
year. The comparison of the gender distri-
bution between Garching and Chile is
shown in Figure 4. The fraction of female
fellows has decreased dramatically in
Chile – by almost a factor of two since
1999, while the distribution in Garching
has been between 25 and 40 % over the
recorded period. Admittedly, very few
applications for ESO Fellowships in Chile
were received from female candidates in
2003 and 2004, nevertheless the trends
show that constant vigilance is required.
The situation with the students is more
positive. While there still appears to be a
majority of male applicants, the selection
has consistently created a more balanced
distribution so that ESO has a fairly high
fraction of female students (see Figure 5),
which is consistently above 30 % and
often above 40 %. The fluctuations are
due to the short-term nature of student
1999
2000
2001
2002
2003
2004
2005
2006
10 20 30 40 50
% %
Figure 4: Percentage of female fellows in Garching,
in Chile, and as a whole.
contracts (two years in almost all cases,
with the International Max-Planck Re-
search School (IMPRS) contracts run-
ning for three years – currently there are
six such positions at ESO Garching). The
distribution between Garching and Chile
is fairly balanced, with the large fluctua-
tions for Chile due to the small number of
students (ten positions).
Overall, the situation with young re-
searchers at ESO has improved over the
past few years. With the exception
of 2003, the ratio of female students has
been consistently above 30 %. The
number of applications from female stu-
dents is also steadily increasing.
We further investigated the composition
of the selection committees for the fel-
lows and the students. In Chile, during
the past few years not a single woman
has been nominated to this committee,
1999
2000
2001
2002
2003
2004
2005
2006
60 70 10
Figure 5: Percentage of female students in Garching,
in Chile, and as a whole.
whereas in Garching, a realistic balance
of women on the committee has been
achieved. The fraction of women on this
committee is higher than the average
of the Faculty astronomers.
c) ESO Governing Bodies and External
Committees
ESO has one main governing body
(Council) and several specialised commit-
tees that have the task of making deci-
sions and/or recommendations on spe-
cific matters. Since membership to these
committees is selected in different ways
(not only among different committees,
but even among different countries for
the same committee), the numbers we
have gathered on this category are more
complex to interpret.
All
Garching
Chile
20 30 40 50 60
We noted that the more technically-
oriented committees show a very low
percentage of female representatives,
whereas other committees (e.g. Finance
Committee and the Users Committee)
have reached a very balanced gender
distribution. However, every committee
shows a positive trend, and the female
representation is increasing, though at a
slow pace.
d) ESO Visitors
The ESO Scientific Visitor Programme
aims to promote the scientific interaction
between ESO and its community and
to enhance ESO’s role as an astronomical
centre of excellence. The programme
is under the responsibility of the Head of
the Office for Science, who appoints the
members of the two Visitor Selection
Committees (VSC, one in Garching and
The Messenger 128 – June 2007 69
Astronomical News
one in Vitacura) and their respective
chairs. Applications to the ESO Scientific
Visitor Programme can be submitted by
any scientist with an interest in ESO ac-
tivities and/or collaboration with ESO staff
members. ESO can also make formal in-
vitations to scientists with a high scientific
profile. On average, ESO Garching is able
to support between four and five visitors
per month.
From the data we have been able to
collect, i.e. the gender distribution of the
applicants, as well as of the approved
and invited visitors over the past eight
years (1999–2006) at ESO Garching, there
is clearly no discrimination in the approval
of the applications – ESO Garching is
usually able to approve almost all applica-
tions. Nevertheless, we note that despite
the fact that the ESO visitor programme
is open to everybody, the number of
female applicants represents a very small
percentage; the reason for which is not
easy to uncover. One possibility is that
for female astronomers with family com-
mitments, it is notoriously difficult to take
leave of absence from their home insti-
tute. However, there is clearly room for
improvement, especially for the targeted
high-profile astronomers: here, the num-
bers are extremely small (only three wom-
en in the last few years), and ESO should
make a special effort to attract senior
women for extended visits.
We note that the composition of the VSC-
Santiago at the beginning of this year
included one female and seven male as-
tronomers, which is identical to the 2006
situation of the VSC-Garching.
e) Invited speakers to Joint Astronomy
Colloquia and ESO seminars
ESO is a very lively scientific environment,
where many scientific talks are organ-
ised on a weekly basis, both in Garching
and in Santiago. ESO-Garching has its
own Lunch Talk (LT) series, organised by
an appointed team of ESO-Garching
staff (usually two to three members, who
change every two to three years). This
committee has the task of running a full
schedule of weekly talks, by inviting
speakers who cover a broad range of sci-
entific activities and interests. Invitations
70 The Messenger 128 – June 2007
Primas F., Status of Women at ESO: a Pilot Study on ESO Staff Gender Distribution
are also made based on suggestions
provided by other ESO staff members.
Together with MPA, MPE and the LMU-
Sternwarte, ESO-Garching also co-orga-
nises the Joint Astronomy Colloquia (JAC)
series, also held on a weekly basis. The
JAC committee includes representatives
from all involved institutes. This is recog-
nised as the main scientific colloquium of
the week.
Table 2 shows that gender distribution in
the Lunch Talk (LT) series is clearly
more balanced than the one among JAC
speakers, but there is clearly room for
improvement here as well. The numbers
in the table are yet another reflection of
the fact that there are more junior than
senior female astronomers. Considering
the higher relevance of the JAC series,
the very low number of female speakers
is discouraging. The JAC committee is
currently chaired by a woman and it will
be worthwhile to monitor the selection
of speakers in the coming years.
Present and future
All the numbers collected so far and dis-
cussed in this report have been taken
at face value, i.e. they represent snapshot
views of the gender distribution among
ESO staff and various ESO governing
bodies and committees, during the last
few years. No attempt has yet been made
to compare these numbers to, for in-
stance, the number of applicants and
eventually short-listed candidates for any
given staff position. This type of com-
parison was readily available for the Fel-
lows and Students only, as discussed
above. Due to this incompleteness, it is
difficult to properly interpret these num-
bers and draw firm conclusions. In what
follows, we attempt to flag the most im-
portant outcomes of this pilot project and
possible future actions that ESO is con-
sidering.
1. Concerning ESO Staff, we have estab-
lished that the gender distribution is
not very balanced (18 % female versus
82 % male). The balance is far from being
satisfactory, especially in the more tech-
nically-oriented divisions, where very few
women are employed. It is disappointing
that this distribution has remained con-
Female | Male
LT JAC
1999 3 | 49 1 | 32
2000 4 | 37 5 | 28
2001 10 | 29 2 | 29
2002 5 | 29 4 | 26
2003 11 | 28 1 | 27
2004 6 | 19 5 | 30
2005 5 | 23 4 | 33
Table 2: Gender distribution among the invited
speakers to Lunch Talks (LT) and the Joint Astron-
omy Colloquium (JAC).
stant for the past eight years. The women
employed at ESO are preferentially in
lower grades with very few women in sen-
ior positions. In other words, the signifi-
cant increase in the number of newly-
hired staff members required by the start
of VLT operations has not resulted in
a corresponding increased fraction of fe-
male staff (at any level).
2. The conspicuous lack of senior female
astronomers is striking. The fraction
of women in the Assistant and Associate
levels is encouraging, although not yet
satisfactory, and it reflects the recent em-
ployment history. The Assistant Astron-
omer level is the entry level for Faculty as-
tronomers at ESO and these are the
junior members. It seems important to
assess what support is required to have
these women succeed both in their re-
search and at ESO, as well as to increase
their number further and ensure that the
fraction does not decline with increasing
seniority. In a few years these numbers
should be critically assessed and the rea-
sons for possible changes examined.
3. The gender distribution among the
ESO Fellows and Students is in general
higher than for the staff and the Faculty.
The selection of the Fellows and Students
does not show clear discrimination, but
constant awareness in the selection
process is still required. The low female
selection rates in 2003 and 2004 are a
warning not to neglect this issue.
4. For the Scientific Visitor Programme, it
would probably be helpful to assemble
a list of high-profile female scientists to
become the target of official invitations to
visit ESO. If no improvement is seen over
a given amount of time, then one should
try to understand the main reasons that
make women decline the invitations (fam-
ily organisational issues, busy profession-
al schedules, etc.).
5. For the Lunch Talks, but especially for
the JAC series, a similar list of high-profile
female scientists could be useful in order
to significantly increase the number of
female speakers. It is well known that the
percentage of female invited speakers at
any scientific conference remains low.
The International Astronomical Union, for
instance, is now checking, for each sym-
posium they sponsor, the gender and
national distribution of the scientific orga-
nising committee and of the invited
speakers, requiring the organisers to pro-
pose alternative names if the gender
balance is not satisfactory. ESO might
want to follow a similar philosophy for its
own series of symposia and workshops.
The importance of this study is that it
provides a baseline for future comparison
and should be used as a reference for
future studies regarding the status of
women at ESO. Once ALMA will have
started early operations and ESO may be
involved in the development of its next
large facility, it may make sense to repeat
this study.
Equal career opportunities require work-
ing conditions that make it possible to re-
concile family needs and career develop-
ment. Being an international organisation
adds extra responsibility to the employer:
it is usually more difficult for expatriates
to fall back on childcare facilities or rely
on family help in the country of employ-
ment other than the home country. ESO
has adopted, and will continue to adopt a
series of measures that aim to contribute
towards an effective personnel policy.
Some changes to the regulations to fur-
ther help women and family members
working at ESO have recently been im-
plemented (e.g. on-site daycare, more
flexibility on maternity leave, as well as
part-time solutions) and discussions of
several more are under way with the ap-
propriate committees.
Independently of these important im-
provements, and despite the fact that the
current situation at ESO closely follows
the numbers of the other EIROforum1 in-
stitutes (CERN, ESA, ILL, EMBL, ESFR,
EFDA), ESO needs to better understand
why the number of female staff has not
increased over the past few years. Is ESO
not attractive enough for female astrono-
mers and engineers? Are the ESO work-
ing conditions not suitable for female staff
to simultaneously cope with professional
career and family commitments? Con-
stant monitoring and increasing aware-
ness will hopefully lead to improvements.
Acknowledgements
The investigating team would like to warmly thank
the following people for their invaluable help and
1
EIROforum is a partnership of Europe’s seven
largest intergovernmental research organisations,
which includes: the European Organization for
Nuclear Research (CERN); the European Fusion
Development Agreement (EFDA); the European
Molecular Biology Laboratory (EMBL); the Euro-
pean Space Agency (ESA); the European Synchro-
tron Radiation Facility (ESFR); the Institute Laue
Langevin (ILL); and ESO.
At the control station
of Melipal (UT3).
prompt support in gathering some of the numbers
presented in this document: Francky Rombout,
Christina Stoffer, Pamela Bristow, Karin Hansen,
Sandro D’Odorico, Fernando Comerón, Andreas
Kaufer and Markus Kissler-Patig.
References
The following list of publications includes recent
results from studies similar to ours (some of which
were mentioned in the text) as well as some sug-
gestions for further reading. These references can be
found in the form of articles or web sites. The list
does not intend to be exhaustive, rather a starting
point for the interested reader.
The Baltimore Charter for Women in Astronomy
(http://www.stsci.edu/stsci/meetings/WiA/
BaltoCharter.html)
The Pasadena Recommendations (http://www.aas.
org/~cswa/Equity_Now_Pasadena.pdf )
The European Commission reviews on Gender
Equality (http://ec.europa.eu/employment_social/
gender_equality/index_en.html ) and their
publications, including the most recent (2007)
report (http://ec.europa.eu/employment_social/
emplweb/gender_equality/publications_en.cfm)
A Study on the Status of Women Faculty in Science
at MIT (1999 and its 2002 update) (http://web.mit.
edu/faculty/reports/pdf/sos.pdf )
Review of the Status of Women at STScI (2002)
(http://www.aura-astronomy.org/nv/womensRe-
port.pdf) and the AURA (Association of Universi-
ties for Research in Astronomy) response to the
report (http://www.aura-astronomy.org/nv/
response.pdf )
The AIP (American Institute of Physics) Report on
“Women in Physics and Astronomy, 2005” (http://
www.aip.org/statistics/trends/reports/women05.
pdf )
The IAU (International Astronomical Union) Working
Group of Women in Astronomy (http://astronomy.
swin.edu.au/IAU-WIAWG/)
The AAS (American Astronomical Society) Commit-
tee on the Status of Women in Astronomy (http://
www.aas.org/~cswa/)
The IUPAP (International Union of Pure and Applied
Physics) Working Group on Women in Physics
(http://www.iupap.org/wg/wip/index.html)
The Messenger 128 – June 2007 71
Astronomical News
Report on the International Workshop on
Observing Planetary Systems
held at ESO, Vitacura, Chile, 5–8 March 2007
Michael Sterzik, Christophe Dumas
(ESO)
Motivation
Nowadays, the ESO premises in Vitacura
host more than 80 PhD students, fel-
lows and astronomers, and represent the
research centre for the scientific staff de-
ployed at the different observatory sites
in Chile. Several topical working groups
help locally to actively promote and foster
joint research initiatives among ESO sci-
entists. One example is the “Planetary
Sciences Research Group” at ESO Chile
(http://www.sc.eso.org/santiago/science/
PlanetaryGroup), which seeks to under-
stand the formation of planetary systems
at large and the place occupied by our
own Solar System. Group members are
actively involved in observation-oriented
programmes making use of ESO facilities
to carry out front-line research ranging
from discoveries of new brown dwarfs
and exoplanets, to the study of primitive
Solar System bodies.
Encouraged by the fruitful interdiscipli-
nary research approach of our own
group, we proposed to gather both com-
munities of Solar System and extra-
planetary system scientists and organise
the international workshop “Observing
Planetary Systems”. The main idea was to
explore the synergy between these two
communities and confront them with four
key topics: from Discs to Planets; Search
for Planets; Planetary Chemistry; Towards
other Earths. In order to establish such
meetings as part of our research culture,
we selected the venue to be in our refur-
bished, large, conference room in Vita-
cura, equipped with state-of-the-art au-
dio-visual facilities.
The response from the community was
overwhelming. While our original plans
were to limit attendance to 60–80 partici-
pants, nearly 120 scientists participated
in the workshop, many of them worldwide
recognised leaders in their field: about
half of the participants were from Euro-
pean countries, 20 % from the USA, 20 %
from Chile, 5 % from other South Ameri-
can countries, and a few researchers
joined us from Japan and Australia. A
healthy number of students (~ 30 % of the
participants) demonstrated the attractive-
72 The Messenger 128 – June 2007
ness of the field for the next generation of
researchers. Each of the four sessions
was organised around three presenta-
tions by invited speakers, each of 45 min,
all contributing to the discussion from a
complementary perspective of Solar Sys-
tem and extraplanetary sciences. The
addition of contributed talks and a large
number of posters deepened our under-
standing of each subject. All available
time-slots were intensively used for lively
and open discussions and some poster
pop-up sessions were organised at the
end of each day.
Scientific highlights
A few subjectively selected highlights
may demonstrate the scientific ideas and
prospects presented during this work-
shop.
The physical, chemical, and morphologi-
cal evolution of circumstellar discs from
gas-dominated rotating optically thick
structures towards thin planetesimal discs
is a key to understand the formation of
planetary systems in general, and our own
Solar System in particular. Disc models
can be best constrained by applying a
combination of modern observing tech-
niques, utilising the highest spatial resolu-
tion imaging in the optical and NIR in
scattered light, polarimetry and SED de-
termination. Dust settling, i.e. the vertical
segregation of particles with different
masses and sizes can now be directly
probed (Menard). The dynamical history
of our own Solar System, as recon-
structed by sophisticated numerical
N‑body integration, teaches us charac-
teristic differences and/or similarities with
known extrasolar systems, such as why
does our Solar System not have a hot
Jupiter? The most likely reason is the res-
onant hierarchical configuration of our
four gas giants. Late heavy bombard-
ment, a cataclysmic episode of planetesi-
mal infall on terrestrial planets, requires a
large reservoir of planetesimals and ap-
pears therefore compatible with the dust
excesses observed in debris discs
(Morbidelli). Accuracy limits of radial ve-
locity searches for exoplanets are con-
tinuously improving, and the physical
limits have still not been reached (con-
trary to what was thought ten years ago).
Precision RV studies will continue to be
of highest relevance for the massive
follow-up of planetary transit candidates
(Queloz).
Direct imaging of extrasolar planets is a
key science driver for many future ground-
and space-based instrument develop-
ments. A few extrasolar giant planets
(EGPs) around nearby young stars are
already in reach of current adaptive-
optics assisted NIR companion search
programmes, and a leap of the astro-
physical understanding of EGPs is ex-
pected with the next-generation high
contrast imagers, like SPHERE (Mouillet).
Astrochemistry provides another im-
portant link from our own Solar System
to other planetary systems. Deuterium-to-
hydrogen ratios appear remarkably simi-
lar in comets and in the interstellar me-
dium (Kamp). Comets, as messengers
from the early Solar System, may reveal
the answer to questions like, why is the
Earth wet and alive? (Mumma). Transiting
(eclipsing) planets play a key role in un-
derstanding their physics and chemistry.
Atmospheres of hot Jupiters can be stud-
ied in detail by transmission spectros-
copy, while mass-radius-composition
relations allow the interior of super-earth
planets to be probed: the era of com-
parative exoplanetology has just started
(Charbonneau)!
The search for signatures of life on exo-
planets by the detection of atmospheric
and surface biomarkers is a far-reaching
goal of future, ambitious, space-based
missions. Both ESA and NASA are active-
ly promoting missions not only to detect,
but also to characterise, the physical
conditions of terrestrial planets (Fridlund,
Lawson). Today, Earth is still the only
known planet that hosts life, and it can
serve as a template to discuss potential
biosignatures in other habitable worlds.
The search for life beyond our Solar Sys-
tem may soon become a reality – exciting
times are ahead (Kaltenegger)!
Contributed talks and posters presented
the latest results from ground-based and
space-based (e.g. Spitzer) observatories
in the search for exoplanets and the
study of planetary-system formation, in-
cluding our own. The prospects for future
ground-based interferometric, radial-
velocity and high-contrast imaging instru-
ments in the field of exoplanet search
Participants at the workshop on
“Observing Stellar Systems”
photographed in the garden at ESO
Vitacura.
were presented, while several contribu-
tions highlighted the importance of com-
bining in situ spacecraft missions of
the planets and small bodies of our Solar
System with ground-based supporting
observations (Cassini/Titan, Rosetta/
Churyumov-Gerasimenko, Mars space
missions). Although we are still lacking
evidence for the presence of life outside
the Earth, the discussions generated dur-
ing this four-day workshop deepened
our belief that the search for signatures of
life in our own Solar System will provide
strong guidance to the future exploration
of exoplanetary systems.
Summary
On purpose we opted not to publish
printed proceedings of this workshop. In-
The Re-launch of the ESO Web
Rein Warmels, Gabriele Zech (ESO)
Recently, the ESO Web went through a
major revision and was re-launched
with a new Look and Feel and new navi-
gation tools. This article gives an over-
view of why and how the ESO Web has
changed.
stead, all contributions, both oral and
poster, were made available for download
immediately after the workshop from the
conference website (http://www.sc.eso.
org/santiago/science/OPSWorkshop) –
given the consent of the authors. This de-
cision was based on our belief that pro-
viding instant access to the research
highlights presented during this meeting
was more useful than waiting for printed
proceedings in a competitive field like this
one, where results are rapidly evolving
and will be published in refereed journals.
The idea to provide near-instantaneous
access to the material discussed in this
workshop was developed even further by
providing, experimentally, a live-stream
webcast during the conference (video-
streams are also available for download
on the conference website). We also be-
lieve that the costs saved have been
The ESO Web plays an integral and in-
dispensable role in the process of doing
science with ESO’s observing and
archive facilities. It provides an effective
and adaptive medium for exchanging
information, documents and images be-
tween scientists, engineers, the media
and the general public. Furthermore, it
provides various services to the commu-
nity of users of ESO’s observing facilities
well invested in making an attractive pro-
gramme, and supporting some partici-
pants to enable their visit in Chile. In total
about 40 % of the participants received
some financial aid; half were students,
half researchers, and all students had
their registration fees waived.
We would like to thank ESO for allocating
the financial support for this workshop,
and the Pontificia Universidad Católica de
Chile its co-sponsoring. Also, we would
like to thank all people who made this
workshop possible, in particular Maria-
Eugenia Gomez, our librarian who acted
as workshop secretary, the ESO-Chile
administration for the logistical sup-
port received, and the students from the
Universidad Católica de Chile, who
helped during the four days of this work-
shop, notably at the time of the registra-
tion. The flawless local organisation,
the highly praised coffee breaks (includ-
ing delicious appetisers), joint cocktails
and conference dinner all contributed to
a warm and friendly atmosphere during
the workshop, remembered by all partici-
pants.
While meetings of this quality and size
take place naturally in many research
institutes in Europe and North America,
ESO Vitacura largely lacks this experi-
ence and we are looking forward to future
workshops of this kind organised within
our premises.
and is critical for coordination and disse-
mination of information, both internal and
external to ESO, in particular in the area
of science and archive operations.
The ESO Web started its service in 1994.
Since then it has expanded rapidly, both
in the amount of information and serv-
ices that are provided as well as in terms
of access rates. Currently, the statistics
The Messenger 128 – June 2007 73
Astronomical News
show that on average 100 000 pages are
viewed per month, representing a data
transfer volume of the order of 20 Gbytes.
In spite of the information and services
that have been added to the ESO Web
continuously over the past years, its struc-
ture has not been adapted; the ‘Look
and Feel’ still reflected the predominantly
science-oriented approach and usage
of the ESO Web in the mid-nineties. Since
then ESO’s observing facilities have been
greatly enlarged: Paranal has become
fully operational, the concept of service
observing was introduced to maximise
the operational efficiency and scientific
productivity, and the ESO/ECF Science
Archive and Virtual Observatory archival
research facilities are being offered. In
addition, over the past years a large num-
ber of science collaborations and pro-
jects, and public outreach programmes,
were initiated.
New information and services were
added arbitrarily to the old structure and
could not easily be integrated into the
existing navigation. Consequently, the
ESO Web became increasingly complex,
barely maintainable and it became in-
creasingly difficult to store new informa-
tion and services that would be easy to
find and use.
In particular, the start of science opera-
tions of the VLT and its scientific re-
sults triggered a substantial increase in
the awareness of ESO amongst the
public and the media and, consequently,
a noticeable change of user profile of
the ESO Web. Meanwhile, nearly every
private citizen has Internet access and
expects that information can be found
easily. To support this new aspect of In-
ternet usage, which includes the require-
ment to present ESO’s activities to the
general public, ESO’s Web presence had
clearly to be revised and improved.
During the past months ESO’s Public Af-
fairs Department (PAD) has been working
on a new ESO Web. This work was done
in collaboration with the IT Department.
It was guided by the conclusions and rec-
ommendations of an internal working
group, taking into account that over time
the overall user profile has changed
significantly. In this sense ESO is follow-
ing other organisations, such as CERN, in
74 The Messenger 128 – June 2007
Warmels R. and Zech G., The Re-launch of the ESO Web
adapting to the fact that the large majori-
ty of web users are from the public do-
main. Other objectives for making a new
ESO web included:
– improve the appearance, i.e. give the
ESO Web a more modern ‘Look and
Feel’;
– improve the navigation and provide
more functionality;
– improve page content, in particular
make sure that the content is up-to-
date. Remove old and outdated pages
and information;
– restructure the site to serve the differ-
ent user groups more efficiently;
– find and assign responsible persons for
well-defined areas.
The new ESO Web has three major user
areas:
1. Public. This area is intended for the
general public, press and media, (po-
tential) industrial partners and people
interested in working at ESO. The area
is completely new, partially based on
existing information, but also with new
content.
2. Science Users. The area is intended for
professional astronomers who are
doing, or are planning to do, research
using ESO facilities.
3. Intranet. This area is for ESO Staff only.
The new Look and Feel has been imple-
mented throughout the whole Public
area (ESO for the Public – see Figure 1)
and includes information of general inter-
est. Examples are descriptions of ESO’s
observing facilities and the astronomy
programmes using these facilities, con-
tact and travel information, ESO’s public
affairs’ activities (including information for
press and media), the new ESO Public
Image Archive (Figure 2), as well as infor-
mation for ESO’s industrial partners and
about current vacancies.
The Science area (Science Users Infor-
mation) provides information and serv-
ices for professional astronomers cover-
ing ESO’s observing and archive facilities.
Examples are instrument information,
Phase I proposal submission, policies
and procedures, service observing and
data processing. Also information about
science activities, science meetings,
the library, and publications can be found
here. In the Science area pages at the
top navigation levels have been imple-
mented with the new Look and Feel.
Lower level pages that are still in the old
format will be converted later.
Most of the information in the Intranet
area is still in the old layout. It is intended
that also in this area the new Look and
Feel will be implemented.
The new structure of the ESO Web is re-
flected in the navigation bar to the left
and the breadcrumb navigation at the top
of each page. The main areas of Public,
Science, and Intranet can be accessed
via separate buttons. Each area has its
own navigation menu that helps the user
to easily find the information or service of
interest.
Figure 1: The new
ESO home page (Public
Area).
The new ESO Web can be reached via
ESO’s standard URL: http://www.eso.org/,
which brings the user to the Public portal.
The Science Users and Intranet naviga-
tion buttons bring users to the Science
and Intranet portals respectively; these
pages can also be bookmarked as start
pages for the ESO Web.
In the current implementation the ESO
Web still consists of many single web
pages that use a common style and nav-
igation. In the near future it is planned to
move to a database-based Content Man-
agement System. This will substantially
improve maintainability of the site and will
serve all web users with the most cur-
rent information and a common Look and
Feel.
Fellows at ESO
Maria Messineo
I have been an ESO fellow since Septem-
ber 2004. I am interested in studying
the morphology and evolution of the
Milky Way. Although it is well established
that our Galaxy is a barred spiral galaxy,
the properties of each Galactic compo-
nent – Disc, Halo, Bulge, central Bar – are
still poorly constrained. It is difficult to
properly map the large-scale morphology
of the Milky Way, mainly because we
observe it from inside the Disc and be-
cause light absorption from interstellar
dust strongly hampers an unbiased view
of its stellar content.
Like Archimedes, I grew up on the an-
cient and wonderful island of Sicily. I
started to study astronomy at the Univer-
sity of Bologna where I graduated in the
spring of the 1997 with a Master’s Thesis
on Galactic globular stellar clusters. I
did my Ph.D. research in Leiden studying
the distribution of stars in the inner re-
gions of our Galaxy as a direct probe of
the gravitational potential.
Maria Messineo
During my Ph.D. I went observing with
the IRAM 30-m telescope in Spain sev-
eral times as well as with the Heinrich
Hertz Telescope in Arizona, with the ESO
3.6-m and the CTIO 4-m telescopes in
Chile. I always enjoyed my time at the
telescopes very much, and so it was ob-
vious that when I joined ESO I would
perform my functional work as a sup-
Figure 2: The new ESO
Public Image Archive.
The page shows
the overview of galaxy
images currently avail-
able.
port astronomer on Paranal. At Paranal I
mostly work with the UT4-Yepun tele-
scope and operate both of the infrared
detectors: NACO and SINFONI. I really
enjoy observing and supporting visiting
astronomers during their runs, and each
time I learn a lot, both scientifically and
technically, and receive valuable feed-
back concerning my research.
Currently I am investigating the spatial
distribution of young stellar clusters in the
Milky Way using SINFONI data. The aim
of this project is to better understand the
current star formation in our Galaxy, as
well as the locations of spiral arms. From
next July I will continue to work on this
topic at the Rochester Institute of Tech-
nology (USA): more science to carry out
and a new world to discover!
The Messenger 128 – June 2007 75
Astronomical News

Laura Parker galaxy evolution and fundamental cos-

I arrived at ESO in late 2005 after com-
pleting my Ph.D. work at the University of
Waterloo in Canada. I am generally inter-
ested in all questions related to the for-
mation and evolution of galaxies, as well
as understanding the fundamental cos-
mological parameters which govern the
evolution of the Universe. My research
focusses on the link between luminous
galaxies and the dark-matter halos in
which they reside. I study the amount and
distribution of dark matter using weak
gravitational lensing, and to date my work
has focussed mainly on galaxy and gal-
axy group sized structures.
Since arriving in Garching I have contin-
ued my work in weak lensing – which is a
leading technique for understanding both
mology. I am enjoying working in such a
stimulating scientific atmosphere, par-
ticularly the many seminars and the lively
discussions at morning coffee. The diver-
sity of research carried out by ESO
researchers is something hard to match
anywhere else. For my functional work I
joined the ESO survey team which over-
sees the planning and execution of public
surveys (for the VST and VISTA). My re-
search makes use of large imaging sur-
veys so this project is a perfect fit to my
interests.
I will be sad to leave Garching (and the
biergartens) behind but I am excited Laura Parker
about moving onto the next chapter in my
career. I will join the faculty at McMaster
University as an assistant professor in
late 2007.

Personnel Movements

Arrivals (1 April– 30 June 2007) Departures (1 April– 30 June 2007)

Europe
Arbogast, Dina (F) Secretary/Assistant
Austin-May, Samantha (GB) Assistant Head of
Personnel Department
Cikic, Zoran (F) Senior Contract Officer
Clare, Richard (GB) Physicist
Guerlet, Thibaut (F) Student
Haimerl, Andreas (D) Electronics Technician
Hussain, Gaitee (GB) Astronomer
Laaksonen, Milla (FIN) Administrative Assistant
Lombardi, Gianluca (I) Student
Mazzoleni, Ruben (I) Appied Scientist
Ngoumou Yewondo, Judith Savina (D) Student
Penker, Thomas (D) Assistant Facility Management
Tanaka, Masayuki (J) Fellow
Testi, Leonardo (I) ALMA European Project Scientist
Chile
Ederoclite, Alessandro (I) Operations Astronomer
Elliott, David (GB) Student
Haguenauer, Pierre (F) VLTI System Engineer
Europe
Bedin, Luigi (I) Fellow
Foellmi, Cedric (CH) Fellow
Huster, Gotthard (D) Mechanical Engineer
Sadibekova, Tatyana (UZ) Student
Vasisht, Gautam (IND) Software Engineer
Chile
Aguilera, Hugo Freddy (RCH) Accounting Officer
Eschwey, Jörg (D) Civil Engineer
Fischman, Nicolas (RCH) Head of Contract and Procurement
Francois, Patrick (RCH) Operations Astronomer
Garcia, Enrique (RCH) Electronics Technician
Hartung, Markus (A) Fellow
Jehin, Emmanuel (B) Operations Astronomer
Mella, Juan Alberto (RCH) Safety Engineer
Shen, Tzu Chiang (CN) Software Engineer
Soto, Ruben (RCH) Software Engineer
Vannier, Martin (F) Fellow
Zagal, Juan (RCH) Software Engineer

76 The Messenger 128 – June 2007

 ESO
European Organisation
for Astronomical
Research in the
Southern Hemisphere
ESO Fellowship Programme 2007/2008
The European Organisation for Astronomical Research in the
Southern Hemisphere awards several postdoctoral fellowships each
year. The goal of these fellowships is to offer young scientists op-
portunities and facilities to enhance their research programmes by
facilitating close contact between young astronomers and the
activities and staff at one of the world’s foremost observatories.
With ALMA becoming operational in a few years, ESO offers ALMA
Fellowships to complement its regular fellowship programme.
Applications by young astronomers with expertise in mm/sub-mm
astronomy are encouraged.
In Garching, the fellowships start with an initial contract of one year
followed by a two-year extension (three years total). The fellows
spend up to 25 % of their time on support or development activities
in the area of instrumentation, operations support, archive/virtual
observatory, VLTI, ALMA, ELT, public affairs or science operations at
the Observatory in Chile. In Chile, the fellowships are granted for
one year initially with an extension of three additional years (four
years total). During the first three years, the fellows are assigned to
one of the operation groups on Paranal, La Silla or APEX/ALMA.
Fellows contribute to the operations at a level of 80 nights per year
at the Observatory and 35 days per year at the Santiago Office. Dur-
ing the fourth year there is no functional work and several options
are provided. The fellow may be hosted by a Chilean institution (and
will thus have access to all telescopes in Chile via the Chilean ob-
serving time). Alternatively, she/he may choose to spend the fourth
year either at ESO’s Astronomy Centre in Santiago, or at the ESO
Headquarters in Garching, or at any institute of astronomy/astro-
physics in an ESO member state.
All fellows have ample opportunities for scientific collaboration within
ESO, both in Garching and Santiago. For more information about
ESO’s astronomical research activities please consult http://www.
eso.org/sci. A list of current ESO staff and fellows and their research
interests can be found at http://www.eso.org/sci/activities/­personnel.
Additionally, the ESO Headquarters in Munich, Germany hosts the
Space Telescope – European Coordinating Facility, and is situated in
the immediate neighbourhood of the Max-Planck Institutes for As-
trophysics and for Extraterrestrial Physics and only a few kilometres
away from the Observatory of the Ludwig-Maximilian University.
In Chile, fellows have the opportunity to collaborate with the rapidly
expanding Chilean astronomical community in a growing partner-
ship.
ESO. Astronomy made in Europe
We offer an attractive remuneration package including a competitive
salary (tax-free), comprehensive social benefits, and provide finan-
cial support for relocating families. Furthermore, an expatriation
allowance as well as some other allowances may be added. The
outline of the terms of service for Fellows (http://www.eso.org/gen-
fac/adm/pers/fellows.html) provides some more details on employ-
ment conditions/benefits.
Candidates will be notified of the results of the selection process
between December 2007 and February 2008. Fellowships begin
between April and October of the year in which they are awarded.
Selected fellows can join ESO only after having completed their
­doctorate.
The closing date for applications is 15 October 2007.
Please apply by filling the web form available at: https://jobs.eso.org
You can provide the following information by uploading the files
directly through the web application form:
– your Curriculum Vitae including a (refereed) publication list
– your proposed research plan (maximum two pages)
– a brief outline of your technical/observational experience
(maximum one page)
In addition three letters of reference from persons familiar with your
scientific work should be sent to ESO electronically (vacancy@eso.
org) before the application deadline. Preferred format for submis-
sions is pdf.
Please also read our list of FAQs at http://www.eso.org/~mrejkuba/
fellows_FAQ.html regarding fellowship applications.
Questions not answered by the above FAQ page can be sent to:
Marina Rejkuba, Tel +49 89 320 06-4 53, Fax +49 89 320 06-4 80,
e-mail: mrejkuba@eso.org
Although recruitment preference will be given to nationals of ESO
member states (members are: Belgium, Denmark, Finland, France,
Germany, Italy, the Netherlands, Portugal, Spain, Sweden, Switzer-
land, the Czech Republic and United Kindom) no nationality is in
principle excluded.
The post is equally open to suitably qualified male and female
applicants.
The Messenger 128 – June 2007 77
Astronomical News
Announcement of
ALMA Community Meeting
3–4 September 2007, Garching, Germany
Now ALMA has entered its main con-
struction phase, ESO and RadioNet are
organising two back-to-back meetings in
Garching, aimed at the European astro-
nomical community.
The ALMA Community Meeting will start
at 14:00 on Monday 3 September 2007
and end in the late afternoon of Tuesday
4 September. The aim of this one-and-a-
half day meeting is to keep the European
astronomical community informed about
ALMA progress since the last ALMA
Community Day in September 2004. The
meeting will provide information about
the project status, and additional reports
on other ALMA activities such as the
status of software. The definition of the
ALMA Operations Plan and the organisa-
tion of the European ALMA Regional
Centre (ARC) has considerably advanced
during the last year. Plans for ALMA op-
eration and for the organisation of the
ARC network in Europe will be presented
and discussed at the Community Meet-
ing, with the aim of obtaining feedback
Announcement of the MPA /ESO/MPE/USM 2007 Joint Astronomy Conference on
Gas Accretion and Star Formation in Galaxies
10 –14 September 2007, Garching, Germany
This meeting will focus on the following
question: how does gas get into galaxies
and what are the processes that regu-
late the rate at which the gas then turns
into stars? The conference will bring
together both theoreticians and observa-
tional astronomers working at different
wavelengths, using different techniques,
both at low and at high redshifts. The
topics to be addressed in the conference
are:
1. H i observations of gas in and around
nearby galaxies
2. The relation between atomic and
molecular gas
3. Insights into the gas-star cycle in
galaxies from new panchromatic data
sets
78 The Messenger 128 – June 2007
Surveys for ALMA Workshop
5–6 September 2007, Garching, Germany
from future ALMA users. Also, a sample
of exciting recent scientific results, rele-
vant to ALMA and the opportunities
afforded by ALMA in its Early Science
phase, will be presented.
The “Surveys for ALMA” workshop will
directly follow the Community Meeting,
starting on Wednesday 5 September, and
end at lunchtime on Thursday 6 Septem-
ber. The rationale behind this meeting is
as follows. While ALMA is being con-
structed in Chile, several ground-based
millimetre and submillimetre observato-
ries worldwide are being upgraded and
are now coming online. The excellent
wide-field survey capabilities of large bo-
lometer arrays such as LABOCA and
SCUBA-2 on single-dish sub-millimetre
telescopes such as APEX and the JCMT,
and the upgraded (sub-)millimetre arrays,
such as IRAM, allow prospective ALMA
users to develop ambitious science
projects and prepare for the use of ALMA.
During the first years of operation of
ALMA, the Herschel Space Observatory
4. Theoretical models and empirical con-
straints on the global efficiency
with which gas is converted into stars
in galaxies
5. Gas inflow mechanisms
6. Feedback processes in galaxies
7. The nature of the Warm/Hot Intergal-
actic Medium (WHIM). Does gas cool
from the hot phase? Insights from
XMM/Chandra
8. Physical constraints on gas in the
vicinity of galaxies from quasar ab-
sorption lines
9. Star formation in high-redshift galaxies
and the Planck mission will also be oper-
ational, and provide unprecedented far-
infrared survey capabilities. The progress
in wide-field near-IR detectors on dedi-
cated telescopes such as VISTA will also
provide a major new data set for ALMA
follow-up observations. The aim of this
one-and-a-half day meeting is to coordi-
nate the planning of these preparatory
surveys for ALMA, and to solicit feedback
from the community in the planning of the
early science follow-up with ALMA from
these surveys. The potential for deep
legacy type surveys with the completed
ALMA array in 2012 will also be briefly
discussed.
If you would like to register for one or
both of these meetings, or would like to
obtain further information, please visit
http://www.eso.org/projects/alma/
science/meetings/gar-sep07/
The registration deadline is 13 July 2007.
Scientific Advisory Committee:
Jacqueline Bergeron (IAP), Andi Burkert
(MPIA), Chris Carilli (NRAO), Francoise
Combes (ObsPM), Andy Fabian (IoA),
Reinhard Genzel (MPE), Ortwin Gerhard
(MPE), Tim Heckman (JHU), Guinevere
Kauffmann (MPA), Rob Kennicutt (IoA),
Eliot Quataert (University of California,
Berkeley), Piero Rosati (ESO), Renzo
Sancisi (INAF), Ken Sembach (STScI),
Mike Shull (University of Colorado,
Boulder), Ian Smail (Durham), Jonathan
Tan (ETH, Zurich)
Further information can be found at
http://www.mpa-garching.mpg.de/
~gassf07/
 Photo: H. H. Heyer, ESO

A field of ‘White Penitents’

– ice sculpted by solar radiation –
at the high altitude of the
ALMA site at Llano de Chajnantor.
ESO is the European Organisation for
Astronomical Research in the Southern
Hemisphere. Whilst the Headquarters
(comprising the scientific, technical and
administrative centre of the organisa-
tion) are located in Garching near
Munich, Germany, ESO operates three
observational sites in the Chilean Ata­-
cama desert. The Very Large Telescope
(VLT), is located on Paranal, a 2 600 m
high mountain south of Antofagasta. At
La Silla, 600 km north of Santiago de
Chile at 2 400 m altitude, ESO operates
several medium-sized optical tele­
scopes. The third site is the 5 000 m
high Llano de Chajnantor, near San
Pedro de Atacama. Here a new submil-
limetre telescope (APEX) is in opera-
tion, and a giant array of 12-m submil-
limetre antennas (ALMA) is under
development. Over 1600 proposals are
made each year for the use of the ESO
telescopes.
The ESO Messenger is published four
times a year: normally in March, June,
September and December. ESO also
publishes Conference Proceedings and
other material connected to its activi-
ties. Press Releases inform the media
about particular events. For further
in­formation, contact the ESO Public
Affairs Department at the following ad-
dress:
ESO Headquarters
Karl-Schwarzschild-Straße 2
85748 Garching bei München
Germany
Phone +49 89 320 06-0
Fax +49 89 320 23 62
information@eso.org
www.eso.org
The ESO Messenger:
Editor: Jeremy R. Walsh
Technical editor: Jutta Boxheimer
Technical assistant: Mafalda Martins
www.eso.org/messenger/
Printed by
Peschke Druck
Schatzbogen 35
81805 München
Germany
© ESO 2007
ISSN 0722-6691
Contents
The Organisation
C. Cesarsky – The Czech Republic Joins ESO 2
J. Palousˇ , P. Hadrava – Astronomy in the Czech Republic 3
Telescopes and Instrumentation
T. Szeifert et al. – FORS1 is getting Blue: New Blue Optimised Detectors
and High Throughput Filters 9
W. Freudling et al. – Towards Precision Photometry with FORS:
A Status Report 13
R. Siebenmorgen et al. – Exploring the Near-Infrared at High Spatial
and Spectral Resolution: First Results from CRIRES Science Verification 17
F. Gonté et al. – The First Active Segmented Mirror at ESO 23
C. Haupt, H. Rykaczewski – Progress of the ALMA Project 25
T. Wilson – ALMA European Project Scientist Appointed 31
Astronomical Science
J.-P. Beaulieu et al. – Hunting for Frozen Super-Earths via Microlensing 33
P. E. Nissen et al. – Sulphur Abundances in Metal-Poor Stars –
First Result from CRIRES Science Verification 38
R. Scarpa et al. – Using Globular Clusters to Test Gravity
in the Weak Acceleration Regime 41
J. Zuther et al. – Dissecting the Nuclear Environment of Mrk 609
with SINFONI – the Starburst-AGN Connection 44
S. Savaglio et al. – GHostS – Gamma-Ray Burst Host Studies 47
C. Tapken et al. – The Puzzle of the Lya Galaxies: New Results from the VLT 51
Astronomical News
M. Grenon – Nature Around the ALMA Site – Part 2 57
U. Grothkopf et al. – Using the h-index to Explore the Scientific Impact
of the VLT 62
F. Primas – Status of Women at ESO: a Pilot Study
on ESO Staff Gender Distribution 67
M. Sterzik, C. Dumas – Report on the International Workshop on
Observing Planetary Systems 72
R. Warmels, G. Zech – The Re-launch of the ESO Web 73
Fellows at ESO – M. Messineo, L. Parker 75
Personnel Movements 76
ESO Fellowship Programme 2007/2008 77
Announcement of ALMA Community Meeting
and Surveys for ALMA Workshop 78
Announcement of the MPA/ESO/MPE/USM 2007 Joint Astronomy Conference
on Gas Accretion and Star Formation in Galaxies 78
Front Cover Picture: The three types of ALMA antenna at the ALMA Test Facility at
Socorro, New Mexico are shown, from left to right: US (Vertex RSI); European (AEC
Consortium); Japanese (Mitsubishi). Since this photograph, the Japanese antenna
has been removed and the US and European antennas have been linked interfero-
metrically (see ESO PR 10/07). Photograph by Hans Hermann Heyer (ESO).